Modelling cherry full bloom using ‘space-for-time’ across climatically diverse growing environments

Highlights

• Knowledge of cherry phenology helps growers manage climate variability and change.

• High quality long-term records are not often available.

• Australian grower records of cherry full bloom were used in lieu of long-term data.

• The sequential and chill overlap models were evaluated, both performed well.

• Data that substitute time-for-space can increase knowledge and improve management.

Abstract

A dataset of cherry full bloom dates across the full diversity of Australian growing regions was compiled and utilised for the first time. The primary data source was from growers located across Australia's major cherry growing districts. Records were of varying length and were used to investigate the potential of using data that substitute ‘space-for-time’ in phenology assessments. Full bloom timing for three cultivars, Bing, Lapins and Van, were collated. The data showed variation in full bloom time across the sites as well as inter-annual variability within sites. This highlighted the potential benefit of a predictive model for growers to better manage the flowering period. The performance of the sequential and the chill overlap models were evaluated using this data. Both models resulted in good statistical representation of the data with RMSE of 3.2 to 5.3 days and 3.3 to 5.4 days for the training and validating data, respectively. The parameterisations of the models differed with the sequential model estimating higher chilling requirement (65–68 CP) than the chill overlap model (46–48 CP). The sequential model estimated similar heat requirements between the cultivars, with Van marginally higher (7250–8000 GDH). The chill overlap model estimated a larger difference in the minimum heat requirement with Van higher (8223 GDH) than Bing and Lapins (6262 and 6486 GDH, respectively). Collation of grower records proved to be a valuable data source which, in the absence of funded programs, could be better utilised for further research that aligns outcomes with industry needs, such as decision support around polliniser varieties, design of growing systems and use of dormancy breakers to synchronise flowering for optimal fruit set and quality.

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.