A MODEL FOR PREDICTING DIAMETER OF 'RED SENSATION' PEARS

Abstract

The pear cultivar ‘Red Sensation’ is showing popularity in the industry, because of its highly-colored, sweet and juicy fruit. Fruit size is critical for marketing pear; therefore, knowledge of growth dynamics becomes essential to program harvesting with minimum losses of size and flavor. The objective of this work was to develop a model to predict the seasonal growth for ‘Red Sensation’ pears, expressed in terms of fruit diameter as a function of time from full bloom. Fruit growth was followed at the Experimental Farm of the Universidad Nacional del Comahue, High Valley region, Rio Negro, Argentina (38°56’S; 67°59’W), located in an arid region, with average annual rainfall of 250 mm, on a sandy loam soil, during the 2005-06, 2006-07, 2008-09, 2009-10 and 2010-11 growing seasons. The orchard was irrigated by surface flooding, and cultural practices were performed according to the local standard program. Trees were selected at random, and maximum fruit diameter (FD) measurements were carried out every two weeks with a Vernier caliper. The range of sampling dates was 19 and 117 days after full bloom (DFB). Equations were developed with SYSTAT procedure. The R values and residual mean squares were used to evaluate the goodness-to-fit of the models. Results showed that the following logistic model provided the most satisfactory fit to the pooled data (n=695), as compared to the power and linear equations: FD (mm)= 81.49/(1 + e ), R=0.92, P<0.001. Fruit maximal absolute growth rate, derived from the selected function, was 0.61 mm/day. A prediction chart was based on the development of the equation and showed ‘Red Sensation’ pear sizes, at various times after 112 DFB, with practical application to aid crop marketing. This model can also be used for planning orchard practices such as thinning and irrigation. INTRODUCTION The pear cultivar ‘Red Sensation’ (Pyrus communis L.) was first raised in Australia in 1940. It is showing popularity in the industry because of its highly-colored, sweet and juicy fruit. Fruit size is critical for marketing pear; therefore, knowledge of growth dynamics becomes essential to program harvesting with minimum losses of size and flavor in this climacteric fruit. Methods for accurate prediction of fruit size and quality attributes are increasingly required as tools for achieving competitive advantages for fresh-marketing services. Winter (1987) reported that the fruit growth curve and the relationship between fruit weight and diameter of each particular cultivar, were essential components in the development of mathematical models. Forecasting methodology should provide estimates with known precision that can be calculated using the smallest sets of easily collected, simple measurements (De Silva et al., 1997). The seasonal course of growth and development is a life process genetically determined, hormonally regulated and modified by location. This indicates that specific fruit growth curves are required according to particular cultivar, soil, climate and orchard management conditions (Ortega et al., 1998). Several factors affect final fruit size, such as cell number, cell size, intercellular spaces and rate of cell division, time and severity of thinning, temperature, irrigation and light stress (Garriz et al., 2009; Alvarez et al., 2010). Different types of seasonal growth curves of pears were reported elsewhere Proc. IX IS on Modelling in Fruit Research and Orchard Management Ed.: G. Bourgeois Acta Hort. 1068, ISHS 2015 148 (Marsal et al., 2000; Arzani et al., 2008; Martins et al., 2008). While pear fruit ripening has been well studied, because of its importance in fruit harvest and storage (Garriz et al., 2008), there is still relatively little information on the growth pattern characteristic of each cultivar. In the High Valley region of Argentina, the fruit growth pattern of ‘Bartlett’ (Garriz et al., 1996), ‘Packham’s Triumph’ (Garriz et al., 1999) and ‘Abbe Fetel’ (Garriz et al., 2004) were determined under similar field conditions. The objective of the present study was to develop a model to predict the seasonal growth for P. communis L., ‘Abbe Fetel’, expressed in terms of fruit diameter, as a function of time from full boom, under the orchard conditions of the Universidad Nacional del Comahue. MATERIALS AND METHODS A crop of ‘Red Sensation’ pear trees on P. communis L. rootstock, planted in 1993, at 4.002.80 m spacing, was studied at the Experimental Farm of the Universidad Nacional del Comahue, in the Rio Negro High Valley region of Argentina (38°56’S; 67°59’W), on a sandy loam soil. Trees were trained to palmette and row orientation was north-south. The orchard was kept weed-free, fertilized, thinned, pruned and sprayed for pest and disease control according to the local standard program for pears. Trees were surface irrigated at weekly intervals to match the crop evapotranspiration requirements throughout the season. The experimental site was located in an arid region, with average annual rainfall of 250 mm. An automated meteorological station (Metos, Gottfried Pessl., Weis, Austria) situated close to the experimental orchard continuously monitored maximum, mean and minimum air temperature during the growth period. As shown in Table 1, the monthly temperature, in our study, varied little between seasons. The range of minimum and maximum temperatures was 1.7-4.1°C and 29.9-31.8°C, respectively. Five trees were selected at random, and four fruits were sampled every two weeks until commercial harvest, during five consecutive growing seasons: 2005-06, 2006-07, 2008-09, 2009-10 and 2010-11. Full blossom was estimated to be at September 26, September 25, September 27, October 5, and September 29 (2005, 2006, 2008, 2009 and 2010, respectively) (Table 2). Maximum fruit diameter (FD) measurements were carried out with a Vernier caliper, to the nearest 0.01 cm (model 30-410-5, General Supply Corporation, Jackson, Miss., USA), and the range of sampling dates was 19 and 117 days after full bloom (DFB). A total of 695 fruits were measured. FD was regressed against DFB. Equations were developed with SYSTAT procedure. Model suitability was evaluated using goodness-to-fit measures. RESULTS AND DISCUSSION Several models were considered to determine the most consistent in being an adequate representation of the data. The R values and residual mean squares were used to evaluate the goodness-to-fit of the equations. It was found that the following logistic regression provided the most satisfactory fit, with the highest coefficient of determination for the pooled data (0.92), compared to the power and linear models (0.91), and the residual mean square from fitting the logistic model was the smallest (Table 3):