Modelling cherry full bloom using ‘space-for-time’ across climatically diverse growing environments
Introduction
Understanding flowering phenology in temperate fruit trees is an area of active research across the globe (Chuine et al., 2016; Darbyshire et al., 2017; Legave et al., 2013; Wenden et al., 2016). This interest has been spurred by potential production impacts from anthropogenic climate change and has been considered for some time (Cannell and Smith, 1986). Flowering phenology in temperate fruit trees is synchronised with seasonal fluctuations in temperature (Saure, 1985) and the timing varies between seasons and growing locations (Richardson et al., 2013). Temperature increase as a result of climate change may alter phenophase timing and influence productivity. Potential productivity limiting factors include asynchrony of cross-pollinating cultivars reducing yields, increased frost risk and reduced winter chill in warming growing regions that decreases synchronicity of flowering (Campoy et al., 2019; Chmielewski et al., 2004; Kaukoranta et al., 2010; Luedeling et al., 2011). Development of reliable models of flowering phenology can contribute to climate adaptation strategies through better characterisation of risks and opportunities and assist industry better manage climate variability through improved prediction.
A key requirement to the development and testing of phenology models is phenological observations to train and validate the model. Ideally, longitudinal data (or panel data) collected consistently would be used to model phenology as this should reduce model error due to differences in tree management and minimise observer error. There are some instances of longitudinal phenological data of substantial length. For example, grape harvest date records in excess of 600 years have been compiled (Daux et al., 2012). However, typically data are collected over shorter periods and may be tied to non-ongoing experiments. For example, Wenden et al. (2016) collated sweet cherry phenology data across Europe and noted variable data length, from 5 years up to 37 years.
Given the reality of data availability, restricting modelling research to longitudinal datasets would limit scientific advancement. As a result, researchers have conducted research by combining data from different sites to construct a larger dataset through using data that substitute space-for-time. For example, phenology modelling by Pope et al. (2014) used almond flowering records of various length from three Californian sites, resulting in a total dataset of 54 records for the period 1984–2008. Similarly, Luedeling et al. (2009) used walnut phenology observations from eight different sites and Legave et al. (2015) used apple flowering observations from 10 sites. Most often these studies utilise researcher collected data from experimental orchards. This view of data acceptability likely restricts broader research into fruit tree flowering phenology, particularly for countries and crops where investment in research orchards is limited.
In Australia there is no long-term government (or other) supported phenology program for temperate fruit trees and experimental orchards are dwindling. For cherry, there is no ongoing researcher collected dataset and few data from previous experiments. As a result, no assessments of cherry flowering phenology in current or future climates have been conducted for Australia. This knoweldge gap has created uncertainty for the cherry industry in terms of strategic investments, such as investment in orchard expansion or selecting cultivars that will be suitable under future climates. This study collated the first Australian dataset of flowering phenology for three cherry cultivars, ‘Bing’, ‘Lapins’ and ‘Van’. The primary data source was grower records from commercial orchards. Without these records no assessment would be possible. Other than availability, the advantage of this data was wide coverage across the main growing regions encapsulating distinct climate regions across Australia. This attribute of capturing data from varied climates to understand flowering phenology is important when assessing prediction models. If phenology models are to be used to inform strategies under climate change these model need to be stable across diverse climates (Richardson et al., 2013). When considering climate variability and the testing of phenological models, a single longitudinal dataset may not contain sufficient variability.
The industry collated cherry flowering phenology dataset enables the investigation of the relationship between phenology and temperature conditions. Various phenology models are available due to uncertainty around physiological mechanisms that control bud burst and flowering (Campoy et al., 2011b; Luedeling, 2012). To further illustrate the potential value of a grower collected database, two flowering phenology models were assessed in this study. These were the commonly used sequential model (Ashcroft et al., 1977) and the chill overlap model (Pope et al., 2014). In their study Pope et al. (2014) found that the chill overlap model predicted flowering time in almond well (Root mean square error – RMSE - of 3.5 to 3.9 days). It was also found to reduce RMSE when compared to the sequential model for Australian and global analysis of apple (Darbyshire et al., 2016, 2017). However, Prats-Llinàs et al. (2018) did not find that the chill overlap model performed better than other phenology models for winegrape data from California and Spain. Model evaluation for an alternate crop, cherry, will test the robustness of the chill overlap model in comparison to the sequential model.
This study aimed to investigate the potential value of grower records of cherry full bloom dates to fill knowledge gaps. The value was demonstrated through 1) quantifying the variability in full bloom timing across Australia's growing districts, 2) assessing two phenology models for predictive capacity and, 3) investigating chilling and heat requirements of the three cultivars.
Section snippets
Methods and materials
Full bloom datasets for three cherry cultivars (‘Bing’, ‘Lapins’ and ‘Van’) were collated from 18 orchards across Australia's major cherry growing regions (Fig. 1).
The sites represent a range of different growing climates in Australia (Fig. 2). Differences in winter temperatures are more pronounced than spring with sites from Western Australia (Kirup and Manjimup) notably warmer than the eastern Australian sites. Mean spring temperatures were more consistent between the sites with the Tasmanian
Results
The variability in full bloom dates for each cultivar highlighted the diversity in Australian growing conditions. Across years and sites full bloom date ranged across 32, 30 and 30 days for Bing, Lapins and Van, respectively (Fig. 3). Within site variability was also apparent with, for example, full bloom for Bing and Van at Ardmona observed across a 25 and 24 day range, respectively.
Using these data the best performing phenology models, statistically, for each cultivar were tabulated for the
Discussion
This study developed and utilised the first dataset of flowering phenology complied for sweet cherry in Australia. The data were primarily sourced from grower records, covered the major cherry growing regions in Australia and included three cultivars (Bing, Lapins and Van). The data were sourced from sites from a wide range of climate conditions, particularly in relation to winter temperatures (Fig. 2). This diversity in climatic growing conditions provided the variability needed to develop and
Conclusions
This study has demonstrated the value of grower observations applied to modelling of full bloom date in sweet cherry. Further that through substituting ‘space-for-time’ insights into Australian flowering conditions can be gained through modelling. Appropriate models, if coupled with weather forecasts, have the potential to assist growers now through better prediction of flowering time to support application of dormancy breakers and management of the flowering period for optimising fruit set and
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We are grateful to growers Alex Turnbull, Antony Spruce, Nick Owen, Mark Chapman, Steven Chapman, Simon Rouget and Terry Martella and researchers Dr Penny Measham, Susan Murphy-White, and Susanna Turpin for providing cherry phenology data. We also thank the two anonymous reviewers for providing comments that greatly improved the quality of this research.
References (47)
- et al.
Dormancy related traits and adaptation of sweet cherry in Northern Africa: a case of study in two Tunisian areas
Sci. Hortic.
(2017) - et al.
High temperatures and time to budbreak in low chill apricot ‘Palsteyn’. Towards a better understanding of chill and heat requirements fulfilment.
Sci. Hortic.
(2011) - et al.
Dormancy in temperate fruit trees in a global warming context: a review
Sci. Hortic.
(2011) - et al.
Climate changes and trends in phenology of fruit trees and field crops in Germany, 1961–2000
Agric. For. Meteorol.
(2004) A unified model for budburst of trees
J. Theor. Biol.
(2000)- et al.
A global evaluation of apple flowering phenology models for climate adaptation
Agric. For. Meteorol.
(2017) - et al.
An evaluation of the chill overlap model to predict flowering time in apple tree
Sci. Hortic.
(2016) - et al.
The temperature dependence of dormancy breaking in plants—mathematical analysis of a two-step model involving a cooperative transition
J. Theor. Biol.
(1987) - et al.
Responses of spring phenology in temperate zone trees to climate warming: a case study of apricot flowering in China
Agric. For. Meteorol.
(2015) - et al.
Modeling the effects of winter environment on dormancy release of Douglas-fir
For. Ecol. Manage.
(2010)
Using spatial interpolation to construct a comprehensive archive of Australian climate data
Environ. Model. Softw.
Climate change impacts on winter chill for temperate fruit and nut production: a review
Sci. Hortic.
Validation of winter chill models using historic records of walnut phenology
Agric. For. Meteorol.
Complexity in chill calculations: a case study in cherries
Sci. Hortic.
Evaluation and fitting of models for determining peach phenological stages at a regional scale
Agric. For. Meteorol.
A biologically based approach to modeling spring phenology in temperate deciduous trees
Agric. For. Meteorol.
Climate change, phenology, and phenological control of vegetation feedbacks to the climate system
Agric. For. Meteorol.
Heat model for pistachio bloom and harvest
Sci. Hortic.
Validation of chill unit and flower bud phenology models for ‘Montmorency’ sour cherry
Acta Hortic.
Climate change impacts on phenology and yields of five broadacre crops at four climatologically distinct locations in Australia
Agric. Syst.
Statistical-method of determining chill unit and growing degree hours requirements for deciduous fruit-trees
HortScience
Yield potential definition of the chilling requirement reveals likely underestimation of the risk of climate change on winter chill accumulation
Int. J. Biometeorol.
Climate warming, spring budburst and forest damage on trees
J. Appl. Ecol.
Cited by (14)
Mapping cherry blossom phenology using a semi-automatic observation system with street level photos
2023, Ecological InformaticsAgroclimatic requirements and adaptation potential to global warming of Spanish cultivars of sweet cherry (Prunus avium L.)
2023, European Journal of AgronomyPrediction of apple first flowering date using daily land surface temperature spatio-temporal reconstruction and machine learning
2022, Computers and Electronics in AgricultureCitation Excerpt :Vegetation phenology is a sensitive and important indicator of vegetation in response to climate change (Li et al., 2021), and accurate knowledge of phenological information is essential to improve our understanding of climate disaster warning and agricultural production management (Fan et al., 2022; Ma et al., 2022). In recent years, an increasing number of studies have focused on analyzing the relationship between flowering phenology of temperate fruit trees and climate change (Darbyshire et al., 2020; Funes et al., 2015; Guo et al., 2019; Pertille et al., 2022), and the results suggest that climate change may increase the uncertainty of flowering phenology in temperate fruit trees (Kim et al., 2022; Li et al., 2021). The first flowering date (FFD) is an important phenological parameter for apple growth and development and a critical phenological period that is closely related to apple yield.
Unusually warm winter seasons may compromise the performance of current phenology models – Predicting bloom dates in young apple trees with PhenoFlex
2022, Agricultural and Forest MeteorologyForecasting regional apple first flowering using the sequential model and gridded meteorological data with spatially optimized calibration
2022, Computers and Electronics in AgricultureCitation Excerpt :Two of the main temperature-driven phenology models for temperate deciduous fruit trees are the sequential model (Ashcroft et al., 1977) and the chill-overlap model (Pope et al., 2014). These two models are usually composed of two sub-models: the dynamic chill model (Fishman et al., 1987) and the growing degree hours (GDH) model (Richardson et al., 1975), both of which have been widely used in flowering phenology models (Darbyshire et al., 2017; Darbyshire et al., 2020; Darbyshire et al., 2016; Pope et al., 2014). For example, Pope et al. (2014) proposed a spring phenology model based on biological parameters and starting values and analyzed the flowering data of three apricot cultivars in California over the past 30 years.