ISHS Acta Horticulturae 1068: IX International Symposium on Modelling in Fruit Research and Orchard Management
USING NON-PARAMETRIC REGRESSION TO MODEL DORMANCY REQUIREMENTS IN ALMONDS |
| Authors: | K. Jarvis-Shean, D. Da Silva, N. Willits, T.M. DeJong |
| Keywords: | bud dormancy, chill, climate change, flowering, Prunus dulcis M. |
| DOI: | 10.17660/ActaHortic.2015.1068.16 |
| Abstract:
Bud dormancy, by which buds must experience a certain amount of chill and heat before blooming or leafing out, affects orchard management and cultivar planting choices, and will be an important component to the effects of climate change on tree crops. Previous work has found that there is a partially compensatory relationship between chill and heat requirements in bud dormancy, as opposed to the common conception of one numeric combination of chill and heat alone resulting in bud-break. To date, this relationship has not been satisfactorily modeled for horticultural crops in a Mediterranean climate. This work aimed to address one of the first questions in defining this relationship – when do chill and heat accumulations start? Previous models have assumed starting dates based on historic conditions, or fitted these dates as model parameters. We have instead worked with a number of non-parametric approaches not previously utilized. This study used bloom timing data for Prunus dulcis ‘Nonpareil’, from three locations spanning the Central Valley of California, over ten years (1996-2005). We first used a fitted spline to help identify the most promising starting dates of chill and heat accumulations. Then, we used LOESS to fit polynomial regression functions based on local neighborhoods of data points. Using the Dynamic Model for chill accumulation starting in October, and the ASYMCUR GDH heat accumulation model starting in January, bloom timing was well modeled in 90% of the site-years. These dates differ from many found in the literature, suggesting pomologists and growers may be miscalculating when dormancy requirements have been met. Our results further support the theoretical model of a partially compensatory relationship between chill and heat towards stimulating bud-break. | |
Download Adobe Acrobat Reader (free software to read PDF files) | |
