It has been widely reported that adequate winter chilling is required for release from dormancy, regular bud break and flowering of many temperate woody species to occur in the subsequent year. This chilling requirement is also a protective measure to prevent woody species from growing in adverse winter conditions, and affects the geographical distribution of woody species (Sherman and Beckman, 2003). Additionally, the levels of winter chilling that are required are variable between and also within species, such as Prunus (Valentini et al., 2004), Vaccinium (Song et al., 2013), and Ribes. Blackcurrant (Ribes nigrum L.) has been identified as a temperate woody fruit species that has a particularly high chilling requirement, and as a consequence recent warm winters in the UK have lowered yields and fruit quality, partly as a result of uneven ripening (Jones et al., 2013). Blackcurrant has been grown commercially in the UK for over 100 years (Brennan, 2008), and the juice processing industry has expanded during the latter part of the 20th century, including in northern areas, due in part to the introduction of new frost-tolerant cultivars (Brennan, 1991). However, increasing concern about the damaging effects of reduced winter chilling in blackcurrant has led to an increased focus by breeders on the production of new cultivars with more resilience to changing climatic conditions.

Major questions remain regarding what constitutes a chilling temperature with suggestions that this varies with species and even cultivar. Such information is critical if we are to account for chill in order to predict the potential effects of climate change, or to optimize selection of varieties for any particular climatic region or climate change scenario. It is convenient to treat dormancy as a sequential process, with a requirement for chilling to satisfy the endodormancy requirement and then a subsequent warming requirement to enable bud break (satisfying ecodormancy; Cesaraccio et al., 2004, 2006). Heat requirement (HR) was considered by Richardson et al. (1974) and Citadin et al. (2001) as a major factor determining the subsequent time of flowering in Prunus spp., although it remains unknown whether heat accumulation for flowering and bud break begins before or after the release of endodormancy.

A wide range of chilling functions or chilling ‘models’ have been proposed to account for the accumulation of chill in different species. The most commonly used functions have been (i) the total number of hours below 7.2°C, (ii) the total number of hours between 0 and 7.2°C, and (iii) the ‘Utah’ units which accumulate chill using a weighting function that emphasizes temperatures between 0 and 7°C. Lantin (1973, 1977) recognized that such units do not well reflect the chilling response of blackcurrants which appear to be more sensitive to lower temperatures and developed an exponential chilling function based on that proposed initially by Bidabe (1967). Harrington et al. (2010) proposed a model based on a minimum number of chilling units (termed critical chilling requirement), below which budburst does not occur, and an optimum chilling requirement, after which additional chilling has no effect on timing of budburst. Clearly, between these two points, there can be various combinations of chilling and heat units that can lead to budburst.

It is feasible in principle to estimate chill requirements from a statistical analysis of bud burst in the field in relation to prevailing weather conditions, by taking data from many years and locations, but this approach tends to be rather unreliable as it is difficult to get a representative range of weather conditions. The approach can be improved by transferring cut shoots to an ‘enabling’ temperature environment so as to eliminate complications in the field arising from the warming requirement. More rigorous data can be obtained by imposing the chill to shoots or to whole plants in controlled temperature chambers at a wide range of chilling temperatures (Sunley et al., 2006). Unfortunately this latter approach does not readily lend itself to evaluation of responses to chill in naturally fluctuating environments, and these fluctuating temperature conditions have become increasingly normal in winters in northern area (Henry, 2008; Luterbacher et al., 2009). Erez and Lavee (1971) demonstrated chilling negation by higher temperatures in peach, and subsequently Erez et al. (1979a,b) showed that the level of negation was highly temperature-dependent, with 8 h of high temperatures negating ca. 16 h of chilling temperatures in a daily cycle. Also, the stage of chilling at which the high temperatures were applied was significant, as has been confirmed in apple by Young (1992).

In this paper we compare the performance of a small number of diverse blackcurrant cultivars in two very contrasting years, characterized by widely disparate levels of chilling at key stages throughout the winter. From this, we have compared the effectiveness of different chill models in explaining the correlations between winter temperature fluctuations and budburst.

The work conducted in this study covered two very contrasting winters in Northern Britain, from one of the coldest winters in recent years in 2012/2013 to undoubtedly one of the warmest in 2013/2014. Such variability of winter conditions is increasingly characteristic of the northern European maritime climate. Interestingly, although there were large differences in the amount of time below freezing, the total number of hours below 7.2°C was fairly similar in both winters. The temperatures in the colder winter were closely approximated by the controlled environment treatment used in the 2007–2008 experiment (Jones et al., 2013) where plants were maintained at 0°C.

Conventionally it is assumed that a single chill unit can be used to discriminate effectively between the chill requirements of different cultivars with different cultivars being separated on the basis of the number of units of chill required for effective flowering. On the assumption that chilling is a prerequisite for budburst and adequate flowering, a convenient method for comparison of the utility of different chill unit calculations is to plot the time course of budburst against chill accumulation: for an appropriate chill unit the lines for any 2 years should be coincident. Comparison of such plots for the two extreme winters indicated that the most suitable model to fit the data for the four cultivars in this study differed with both cultivar and year, highlighting the fact that none of the existing models are well-suited to variable field conditions. These results in themselves indicate that adherence to models based on simple temperature thresholds is unlikely to relate to field observations across a range of germplasm in different years.

The need for further development and refinement of models for winter chilling in woody plant species is linked to the increasing temperatures predicted through climate change as summarized by the International Panel on Climate Change (Solomon et al., 2007; Stocker et al., 2013). Indeed, in many areas including California and the UK, the most obvious trends in overall global warming are seen in the daily minimum temperatures during the winter months (Sunley et al., 2006; LaDochy et al., 2007). The overall impacts on the dormancy and development of woody perennial species are inevitably complex and require modeling that can take into account aspects such as timing of warming during the dormant period as well as overall temperature means. Luedeling et al. (2009), comparing different models of winter chill, advocated the use of a dynamic model based on the work of Erez et al. (1990) and Erez and Fishman (1998), and questioned the biological significance of models based on chilling hours alone. Unfortunately effective parameterisation of such models for any species requires data for many years. The fact that some genotypes were relatively little separated by natural chilling in the two experimental years (even when eliminating confounding effects due to the ecodormancy phase by warming in controlled environments) supports the hypothesis that chilling is not the only environmental factor affecting budburst as has been found by previous workers for other woody species (Heide, 1974, 1993), with daylength variation being a prime candidate as another relevant factor.

As climate change may lead to a delay in the onset of chill accumulation, this may also affect the time at which perennial plants become receptive to heat during spring (Luedeling et al., 2013). This is partly borne out by the observations of budburst and subsequent development observed in this study: Table 1 shows that the dates of actual budburst observed in the field averaged 2 weeks earlier for all cultivars in the warm winter of 2013–2014 than in the exceptionally cold winter of 2012–2013. We hypothesize that at least part of the delay in the cold winter may not relate to chill satisfaction, but to fact that the warming phase was delayed by the very late spring in that year. This conclusion is supported by the different results observed when controlled warming was provided. Although flowering was advanced in all four cultivars after the warmer winter of 2013/2014, the genotypic differences in flowering date in the field were largely maintained over the 2 years. The differing responses of the four cultivars used may in part be explained by their relative genetic backgrounds. In particular, the New Zealand cultivar ‘Murchison’ is derived from a low-chill environment, and requires significantly less winter chilling overall than a high-chill Scottish cultivar such as ‘Ben Dorain.’ However, the differences in time of bud break between the cultivars was considerably less following the warmer winter than the colder one, although by flowering this trend was reversed. Despite the genetic differences, the work by Harrington et al. (2010) in pine suggests that the biochemical factors involved in the sensing of chilling are likely to be similar across many species, but the development of suitable species-specific models is still required to accurately quantify some aspects.

Phenotyping of the plant material has to date focused mainly on the chilling temperatures between October and March, with relatively less attention on the effects of warming temperatures occurring both during endodormancy and also as the plants move toward bud break. Such information will be needed to improve the accuracy and applicability of models of chilling and bud break, and to allow wider phenotyping to be undertaken. A particularly useful approach will involve the screening of material grown at different locations (Cesaraccio et al., 2001, 2004). Whether future research should attempt to fit final flowering or the full seasonal dynamics into the model depends on the eventual use of the model, as the latter may be considered of less relevance to plant breeders and to climate scientists, only being of specific relevance to physiologists interested in the mechanism of endodormancy.

Interestingly, it was found by Castède et al. (2014) in Prunus that while the correlation between chilling requirement and flowering date is significant and constant, the correlation with HRs showed a much higher variability between years. There would therefore seem to be a strong case for the provision of more detailed and multi-site phenotyping with particular focus on the impact of warmer temperatures at key points within the chilling accumulation phase, since the work of Erez et al. (1979a) demonstrated the significance of the timing of warming temperatures.

Work to understand the underlying genetics linked to chilling and dormancy release has been reported in Prunus spp. (Sanchez-Perez et al., 2012; Castède et al., 2014). Similar work is in progress in blackcurrant, using both individual cultivars and also mapping populations and genomic resources developed in Ribes (Hedley et al., 2010). In blackcurrant, the availability of significant diversity of response to chilling and warming temperatures during the winter gives some optimism that the application of contemporary genomics and molecular breeding techniques can accelerate the production of new cultivars with resilience to future climate scenarios. However, such work is dependent on accurate and relevant phenotyping, and the future development of more robust and dynamic models for the analysis of responses to winter chill and dormancy release is an essential part of this process. The work both to develop robust chilling models and also to understand the genetic control of chilling accumulation and bud break is not only of considerable significance to the future sustainability of blackcurrant production in Europe, but it also has generic relevance across a range of woody species, both crop and landscape, in the context of a changing climate.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.