Regresar

A regression analysis on factors affecting yield of highbush blueberries

Abstract:

Based on data from a 2012 survey of 189 blueberry growers managing 2,864. ha in the Maule Region (Chile), a regression model was generated to determine the variables that have the greatest influence on blueberry yield. Growers are facing increasing reductions in profitability and yield has a large impact on economical returns. For this, plots from blueberry fields were designated either as high (greater than average yield: 10.1. ton/ha) or low yielding (<10.1. ton/ha). In this study two regression models were estimated: Poisson and logit binomial. The discrete dependent variable was yield (1. =. high yield, 2. =. low yield). The model adequately pbkp_redicted in 73.8% of the cases whether a field was high or low yielding. The averages for yield and plant age for 189 blueberry fields in the survey were 10.1. ton/ha and 6.83-year-old, respectively. Average highbush blueberry yields increase with age up to the 11- to 15-year-old range; as they passed this age, their yields decrease. The average distance of the pickers hired to harvest the fruit was 20. km. The most important highbush blueberry cultivars were Duke (256.7. ha), O'Neal (192.3. ha), and Brigitta (183.6. ha), and they contributed 74% of the total volume of blueberry produced by the region. Most fields (66%) were 6- to 10-year-old. Fields older than 20-year-old were only 0.2% of planted area. Rabbiteye blueberries amounted to 8.8% of total hectarage. The model showed a higher probability of obtaining a high yield with the following conditions: older fields or plants with ages up to 14.7-year-old), pickers coming from a greater distance (inflexion point between 38.3 and 53. km), planting design that included pollenizers, use of mulch to cover soil within the row, weeds controlled with a mixed system, and field management using a conventional production system (versus organic production).