Heat requirements for loquat fruit development may be assessed with a Beta model approach

Abstract

In Mediterranean areas, loquat trees (Eriobotrya japonica Lindl.) typically bloom in fall and fruits develop during the winter period, reaching their ripening in spring. So far, little interest has been driven towards models capable of estimating heat requirements for the accomplishment of fruit development in the winter period. In this work, we verified whether in Mediterranean areas the heat accumulation (GDH) needed for loquat fruit growth, may be precisely described using a Beta model, a flexible mathematical function which, in the past, has been successfully used in peach. BBCH phenological stages of ‘Algerie’ trees were recorded in two experimental sites: Palermo, Italy (38°04’N, 13°22’E, 150 m a.s.l.), and Callosa d’En Sarria, Alicante, Spain (38°39’N, 00°07’W, 150 m a.s.l.). The original data pool included maximum and minimum daily temperatures from a total of sixteen years, nine from Palermo and seven from Callosa. Firstly, the model was tested taking into account phenological data from all the years. However, due to anomalous data, some years were excluded. Temperatures and phenological data – Julian days (JD) from anthesis to ripening date – were used as input to run a Visual Basic computer program for the computation of growing degree hours (GDH). The program was used to randomly change the triplet of cardinal temperatures (i.e., base (Tb), optimal (To) and critical (Tc)) used to construct the Beta model. More than a thousand random sets of cardinal temperatures were generated to optimize the model. The target was to minimize the coefficients of variation (CV) of harvest dates (JD) among years and locations. The best resulting model, having Tb=4.8°C, To=9.0°C and Tc=42.2°C, yielded an average accumulation of 67542 GDH, and CV=2.36%. On average, the model generated an error of 5 days (SE ±1) between real harvest time and forecasts. Tb and To were selected within very narrow intervals, implying a high level of estimation accuracy. Thus, in future studies further curvilinear models have to be studied in order to estimate loquat heat requirements with minimal errors. In conclusion, although further attempts must be carried out in order to refine a loquat fruit growth phenoclimatic model, the approach used in this work may represent a further step to predict loquat harvest date.