Changing times: Opportunities for altering winter wheat phenology

photoperiod-sensitive winter varieties which have a relatively narrow window of flowering time. Changing seasonal conditions alongside mounting logistics pressure as a effective seasonal weather that greater flexibility in varietal flowering time is likely to sustain UK wheat productivity. In this Opinion piece, we present findings from a recent dialog with UK agronomists on the demand and likely potential for greater flowering time range in winter wheat varieties. This provides important insight connecting farmer priorities and demands with ongoing genetic research and breeding. Our findings highlight the farm-level importance of wheat phenology and provide recommendations for future research and plant breeding.


| INTRODUC TI ON
Understanding the sensitivity of production systems to fluctuating seasonal conditions is increasingly important with global warming of up to 4°C above current temperatures predicted by the end of the century (IPCC, 2013). In the United Kingdom, hot summers such as recently experienced are predicted to have a 50% occurrence by mid-century (UKCP18, 2019). On a seasonal basis, greater frequency and severity of adverse weather events will present a significant challenge to maintaining the UK's current level of wheat productivity as is evident from early indications of severely reduced wheat yields in 2020 (National Farmers Union, 2020).
Examples of adverse conditions caused by climate change that may affect wheat productivity in the UK include drought stress, water logging, and adverse conditions at sowing and harvesting (Harkness et al., 2020;Trnka et al., 2014). Given that 40% of UK arable land area is cropped to wheat (DEFRA, 2018), this is likely to have a major impact on both primary production and food supply.
Such impacts on wheat production were observed in the UK in 2012 which was dominated by significant deviation from typical climatic patterns , recording the second driest January to March period since 1953 (for England and Wales; Parry et al., 2013) followed by the wettest 9-month period in 250 years Parry et al., 2013). This had serious implications for winter wheat productivity due to the combination of early season moisture deficit, low sunlight levels during grain fill, and elevated disease pressure associated with high rainfall levels leading to 2012 UK wheat yields suffering a 14% reduction compared to the previous year (DEFRA, 2012).
Flowering time marks the transition from vegetative to reproductive growth and its timing relative to local environmental conditions is the major determinant of grain yield in wheat (Cleland et al., 2007;Flohr et al., 2017;Jung & Muller, 2009). Varieties should ideally flower at an appropriate time not only to maximize radiation intake by photosynthetic tissues, but also to avoid key developmental stages intercepting with adverse abiotic and biotic stresses (Bentley et al., 2013). However, there has been little investigation to date of the interaction between flowering time, ambient temperature, and drought response despite the published predictions of global warming. Wheat is a long-day species with floral initiation accelerated by exposure to lengthening days. At present, regional adaptive response is largely achieved through designed combinations of major genes, namely Vernalization (Vrn-1) and Photoperiod-1 (Ppd-1), which control the transition to flowering in response to environmental cues. Reliance on known genetic combinations to adapt wheat to climate change is an escape strategy: for example, photoperiod-insensitive varieties (possessing mutant Ppd-1a allele/s) are favored in Mediterranean environments where early flowering avoids late summer heat and drought stress (Worland, 1996;Bennett et al., 2012), thereby minimizing yield penalty.
Variation in the major Ppd-1 and Vrn-1 genes aided the worldwide spread of wheat cultivation and the expansion of cultivation into new areas. Olmstead and Rhode (2011) describe the progression of wheat cultivation in North America showing shifts across the climatic features of latitude, longitude, annual and average January and July temperature, and annual precipitation. The development of the early flowering Canadian cultivar 'Marquis' was attributed as a contributing factor in the settlement of western Canada (Morrison, 2008).
The result of a cross between 'Red Fife' and the very early variety 'Hard Red Calcutta' (believed to have been a Himalayan variety sent to foundation Canadian wheat breeder William Saunders by Lord Dufferin, a former viceroy of India; Morrison, 2008), the variety had superior milling quality. Its distinctive earliness, being 6-10 days earlier maturing than the conventional varieties of the time, went on to dominate Canadian wheat production (Olmstead & Rhode, 2011).
'Marquis', along with the early breeding and selection programs in other parts of North America introduced greater environmental adaptation, supporting expansion in wheat production area to match the climatic conditions of the regions (Olmstead & Rhode, 2011).
Similarly, in Asia, the Mediterranean and North Africa the introduction of photoperiod insensitive alleles into improved cultivars translated into enhanced adaptation and higher yield potentials (Ortiz Ferrera et al., 1998), while in Europe photoperiod response was differentially selected between southern and northern production environments with early flowering varieties giving a 33% yield increase (compared to late flowering lines) in southern Europe (Worland et al., 1996;Worland et al., 1998). In China, Yang et al. (2009)  Flowering time has clearly facilitated regional adaptation of wheat and allowed it to become a successful and important crop.
Flowering time is thus a critical consideration in the adaptation of wheat to predicted climate change. Although much can be learnt from historical patterns and drivers of wheat area expansion, future adaptation needs must also be forecast well in advance given the substantial time lag in the breeding and release of new cultivars. Hammer et al. (2020) estimated that adaptative requirements should be considered at least 10 years ahead of varietal release. Based on climate model predictions, wheat yields will suffer climate change-related declines below current production in most regions in the absence of either varietal or infrastructure (e.g. irrigation) interventions (Tanaka et al., 2015). The magnitude of climate change effects on wheat will be cultivar dependent, necessitating practical solutions to tailor selective breeding to changing regional patterns (Trnka et al., 2014). These intervention needs and outcomes can be modeled via so-called "adaptation pathways" but the availability and field-based performance of theoretical future varieties within the required intervention timescales is unknown (Tanaka et al., 2015). Winter wheat is both the dominant cereal and main arable winter crop grown in the UK. Spring wheat, which lacks a vernalization requirement and is sown in late winter or early spring (Figure 1), is also grown but represents a much smaller proportion of the total amount of wheat grown in the UK. Winter wheat varieties in the UK are photoperiod sensitive, most possessing the photoperiod-sensitive allele of Ppd-D1, and therefore are typically later flowering than varieties in other regions of northern Europe (Bentley et al., 2014). Considering the predicted effects of climate change in the UK, breeding varieties that flower earlier and thus pursue an escape strategy like that found in southern parts of Europe, seems a logical way forward.
Here we report on a recent consultation with UK-based agronomists across major wheat producing regions conducted to assess demand for earlier flowering varieties, and to understand associated risks. We define early flowering as a range of variation that is 5-10 days earlier than current UK winter wheat varieties. In this Opinion, we share the findings which indicate that earlier flowering time in UK winter wheat is desirable given more fluctuating seasonal conditions, and that it can mitigate drought stress, contribute to improving the logistics of on-farm crop management, and avoid pest and pathogen damage. We also highlight where additional research is needed, particularly in understanding the physiological link between flowering and maturity date, understanding frost risks, and providing new genetic controllers of adaptive response.

| HARVE S TING VIE WS ON ALTERING WHE AT PHENOLOGY
In order to understand producer perspectives on the opportunities and potential risks associated with altering winter wheat phenology in UK-registered varieties, we consulted agronomists working F I G U R E 1 Schematic representation of the UK winter and spring wheat production window showing the key developmental stages and cropping activities, as well as photoperiod extension along with the biotic and abiotic stresses discussed in the text. Striped regions indicate other possible times of drilling (sowing) or harvest. Red vertical bars indicate the stage of the emergence of the ear (heading), which is the stage with the maximum susceptibility to frost. The developmental timings of the winter wheat reference variety are derived from the Agriculture and Horticulture Development Board (AHDB) Wheat Growth Guide (AHDB, 2018) which used the variety 'Consort', and the timings of the spring wheat variety is based on consultation with agronomists and breeders. Note that developmental stages and subsequently harvest will be later in the north of the United Kingdom due to cooler temperatures. BYDV, barley yellow dwarf virus; OWBM, orange wheat blossom midge directly with farmers across the UK's main wheat-growing regions. In total, we spoke to 14 agronomists, including five from the NIAB regional agronomists' team, four from the NIAB-TAG consulting team, and five from Velcourt Ltd., a farm management and advisory business. Interviewees were selected in order to be representative of regions across the UK and roughly proportionate to the volume of wheat production in these regions ( Figure 2). As only a very small proportion of the UK wheat crop is grown in the North-West of The consultation did not take the form of a formal designed survey but was conducted to ascertain views on future phenology-based research and breeding opportunities.

| MA JOR P OTENTIAL B ENEFITS OF E ARLIER FLOWERING TIME
Across the consultation, a consistent set of opportunities were identified by all the respondents surveyed. These include the provision of a greater range of flowering times to spread seasonal risks, mitigate drought stress, and improve on-farm logistics. All these potential benefits are discussed in more detail below, along with a perspective on likely yield and overall productivity impacts. F I G U R E 2 Wheat production across different regions of the United Kingdom in 2019 with percentage values overlaid to indicate the representation of agronomists consulted in this study. Wheat production data were taken from the Department for Environment, Food and Rural Affairs June Survey of Agriculture and Horticulture (DEFRA, 2019). As wheat production is extremely low compared to the rest of the United Kingdom in Wales (1.2%), Northern Ireland (0.4%), and in North-West and Merseyside (abbreviated to North-West) in England (1.4%), these areas (shown in gray) were not considered in this study 3.1 | Winter wheat varieties with earlier flowering time are desirable given more fluctuating seasonal conditions Overall, there was a very positive response toward having a greater range of varietal flowering times in winter wheat. While all agronomists acknowledged that the demand for early flowering varieties will differ depending on the region due to climate and, within regions, on soil type, they highlighted that there is currently a poor 'JB Diego' is classified as an "early to mature" variety (Senova Limited, 2017); therefore, we classified any varieties having equal or earlier ripening as "early ripening/maturing" and any varieties having later ripening as "late ripening/maturing." We acknowledge that the differences between varieties that are "early" and "late" and are close to the reference do not differ from one another substantially, but we use this classification as a simple illustrative way to show the smaller proportion of "early" varieties especially in more recent years. There was a significant difference in days to ripening across the 10 years tested (p < .001) with a clear decrease in variance of days to ripening (Table S1). Analysis of the AHDB RL thus gives some insight into the changing phenology profile of UK wheat varieties, supporting the assertions made by the agronomists.

| Improving the logistics of crop management is likely possible with earlier flowering varieties
The most commonly acknowledged benefit of early varieties among agronomists was that related to farm and crop management. Farming practices have become more compact with a narrower time window for practical operations placing higher demands on machinery and labor during key cropping activities. Treatment timings, such as ear wash sprays, can be very tight and, at harvest time, there is an upper limit on combine harvesting and significant potential loss in grain quality due to delays in harvesting. Drilling (sowing) is a key pressure point and the drill window is contracting in most regions in the UK in order to drill as much of the crop as late as possible to reduce disease risk and control black-grass growth (summarized in Figure 1). This For each year, ripening (used as a proxy for flowering time) was compared to a reference variety, 'JB Diego'. In the boxplots, the hinges show the first and third quartiles, the middle line is the median value, and whiskers correspond to data points no more than 1.5× the interquartile range from the hinges. Data points each represent an individual variety and are coloured according to the following classification: yellow for late (>0); blue for early (≤0); and the proportion of early ripening varieties is given below the box plots in blue (except for 2017, for which original ripening values were not available and only data rounded to the nearest integer). (b) Winter wheat varieties on the RL over the years 2010/2011 to 2020/2021 show less variation in time to Zadok's growth stage 31 (GS31; start of stem extension) when sown late in the drill window (November) than when sown early (September). The variance for each month is shown above the box plots Cover crops are also designed to return soil health benefits and can include legume, brassica, grass and cereals individually or in mixtures that are generally grown for a longer period between arable crops to improve soil organic matter (AHDB, 2015). This is seen as particularly advantageous in the north of the UK (north of Yorkshire and the Humber) where crops take longer to develop due to cooler temperatures, and early flowering varieties thus provide an assurance that the grain can be harvested before precipitation as well as provide for a catch and cover crop.

| Earlier flowering time can mitigate drought stress
Drought stress is already seen to be a problem in some regions of the UK, particularly on sandy and gravelly soil types. These soil types have a high sand/gravel to clay ratio and do not hold water very well, thereby enhancing any effects of low precipitation. Almost all the agronomists consulted reported that they had worked with farmers that had seen evidence of drought stress, or yield loss due to low water availability and all predict this to continue to be a problem going forward. They could thus see the advantages of wheat with earlier flowering time used to mitigate risk of drought stress.
However, there was also some skepticism that forward weather  with stress events. Therefore, we would argue that from a logistics and infrastructure perspective it will be far more difficult for farmers to mitigate future drought (e.g., enabling use of supplementary irrigation) compared to heat stress (staggering planting dates or using flowering time varietal profiles). European trial environments; Zadoks et al., 1974) and maturity assessed as an average field-based score across three dates (method described and data available in Ladejobi et al. (2019)). Therefore, in these discussions, earlier flowering time was assumed to mean earlier maturation time and a shorter grain fill period, thus leading to the assumption of a yield penalty compared to later maturing varieties, as has been demonstrated in previous investigations of Ppd-1 (Worland, 1996;Bentley et al., 2013). Overall, it was agreed that for growers the variety could not be seen to suffer from too much of a yield penalty.

| Yield will determine uptake of early flowering varieties and greater understanding of the physiological link between flowering, maturity, and the consequences for yield formation is required
Some agronomists discussed the yield advantage that a variety might have if it was early flowering but retained a similar maturation period to later flowering varieties. The major barrier to implementing this manipulation in wheat breeding remains a lack of knowledge relating to the independence of stages of development and their subsequent influence on yield. It has been proposed that photoperiod sensitivity at different developmental phases is at least partially independent (Gonzalez et al. 2005), raising potential for genetic manipulation. Some authors have shown patterns of development differing between photoperiod sensitive and insensitive types (Dyck et al. 2004;Gonzalez et al. 2005;Worland, 1996), while others have reported specific stages (e.g., stem elongation; Tanio and Kato, 2007) to be unaffected by either photoperiod, or genotype.
In a recent simulation study, Harkness et al. (2020)  trade-offs to physiologically based plant processes and productivity across environments need to be accurately characterized. While future wheat ideotypes can be designed and simulated in silico based on climate model predictions (Semenov & Stratonovitch, 2013), it is essential that empirical testing be used to determine if it is feasible to decouple flowering and maturation, and what the downstream production impacts of this would be.

| FURTHER CON S IDER ATI ON S AND RIS K S
In addition to the overriding opportunities discussed above, the agronomists we spoke to raised additional considerations for earlier flowering varieties, including risks associated with pests and pathogens, and frost damage. In addition, we solicited views on dramatically altering phenology to move from winter to spring wheat cultivation although there was little support for this at present.

| Early flowering could reduce the risks of pest and pathogen damage
One potential benefit of earlier flowering is reduced risk of some diseases and pests that affect wheat at flowering, particularly those that infect later in the season. Two in particular were mentioned by agronomists-the ergot fungus, Claviceps purpurea, and orange wheat blossom midge (OWBM; Sitodiplosis mosellana).
Ergot infects the ear at flowering with risk tending to be higher when the weather is wetter and warmer (Gordon et al., 2015).
Timing of flowering is therefore likely to be a major determinant of infection if it co-occurs with these seasonal conditions that favor pathogen dispersal and infection. Wheat does show differential varietal resistance to ergot; however the cause of the variation remains largely uncharacterized (Menzies, 2004;Willis, 1953). Gordon et al. (2015) showed that ergot resistance in UK winter wheat is conferred by at least four quantitative trait loci (QTL), two of which are co-located with the major height reducing, DELLA encoding Rht genes (Rht-B1 and Rht-D1; Peng et al., 1999) with shorter lines with either the Rht-B1b or Rht-D1b semi-dwarfing allele having significantly lower levels of infection (Gordon et al., 2015). In addition to their role in stem elongation, gibberellins are known to control floral development (Cheng et al., 2004); so it is possible that there are pleiotropic interactions between height and flowering time which determine ergot response. Gordon et al. (2015)  For orange wheat blossom midge (OWBM; Sitodiplosis mosellana), which feeds on developing grains, there is a precise window of 10 days in which infection risk is extremely high. Currently, this coincides with flowering of many UK winter wheat varieties that do not carry the Sm1 antibiosis resistance gene (Kassa et al., 2016).
Shifting flowering 7-10 days earlier could help to avoid the peak infection window thus reducing damage from this pest. However, there is anecdotal evidence that earlier flowering varieties previously grown in the UK (e.g., 'Cordiale') have high OWBM infection levels so further detailed characterization is required. Given the narrow genetic basis of current OWBM resistance (Zhang et al., 2020), the availability of a wider selection of varietal flowering time ranges may support the varietal blending approach developed and deployed to counter new virulences in Canada (Smith et al., 2014). Combined with deployment of additional genetic sources of resistance (Zhang et al., 2020) this should provide a more robust resistance package for UK farmers.
While these potential benefits in relation to pests and pathogens were recognized among the agronomists consulted, there was also acknowledgement that changing climates are likely to affect the life cycle of pests and pathogens. The potential and likelihoods for climate change-related effects on the frequency and occurrence of wheat pathogens has been summarized and reviewed (e.g., Chakraborty & Newton, 2011;Juroszek & von Tiedemann, 2013). Therefore, any predicted reductions in risk will need to be combined with modeling and assessment of these factors.

| Increasing the risk of frost damage
Another potential risk identified and discussed was that of damage to earlier flowering varieties from late frosts. Late frosts in spring (and into summer) can harm floral organs and developing grains leading to anther and embryo death and non-fertilization as well as grain damage including small, shriveled, and shrunken kernels leading to an overall reduction in grain production (Cromey et al., 1998). The photoperiod-insensitive Ppd-D1a variety 'Soissons' was discussed by some interviewees in this context, as it was known to be particularly subject to late frosts in UK conditions. It was also acknowledged that the risk of frost increases markedly in the north, one region which could otherwise benefit significantly from earlier flowering (see Section 3.2). Martino and Abbate (2019) developed a model allowing for estimation of the frost damage effect on subsequent wheat grain number across reproductive development stages showing that maximum damage (in respect to grain production) always included anthesis, indicating this as the most critical period for frost-avoidance. Post-head-emergence frost damage is a greater risk in varieties with faster development (recently reviewed by Frederiks et al. (2015)). However, modeling of crop damage risks in the UK from frosts under future climate projections by Trnka et al., (2014) showed that severe frost risk is extremely low with more recent work from Harkness et al. (2020) also reporting that future risks of late frosts leading to medium to severe yield losses across the UK's wheat growing area are negligible. With these current models, it appears that frost may not pose a major risk to early-flowering varieties in the future but, as with any approach, this will require reconsideration as new data becomes available.

| Dramatically altering phenology currently offers limited appeal
We asked the agronomists about the potential of spring wheat and whether they were likely to recommend spring wheat varieties to the farmers that they work with. Of the responses, some were favorable while others were very negative. It was widely acknowledged that a large amount of spring wheat would be planted in the 2019/20 season due to the extremely wet autumn which reduced the sowing of winter wheat, so much so that there was a shortage of spring wheat seed leading to the sourcing of seeds from other parts of Europe. However, a lot of farmers have chosen instead to go with spring barley, which is seen to have more favorable properties compared to spring wheat including more reliable yields, better competitiveness against blackgrass, and lower pathogen and pest issues with spring wheat being especially susceptible to infection by ergot (Menzies & Turkington, 2014) and gout fly (Chlorops pumilionis; Kaniuczak, 2008). In addition, spring wheat must be drilled early (March, or even February; Figure 1) which is not possible due to weather conditions in some regions and on some soil types. There are relatively few options available in terms of spring wheat varieties with only eight spring wheat varieties on the current RL (AHDB, 2020a) and a corresponding low UK certified seed area (1.6% of the total 2019 wheat certified area; NIAB TAG, 2020). Some interviewees were dismissive about the importance of UK spring wheat for the future and most thought that breeders' efforts were better concentrated on improving winter wheat varieties.

| FUTURE PROS PEC TS
Early flowering winter wheat varieties are clearly seen as a desirable option for UK farmers, particularly considering the limited choice that is currently available. The main advantage associated with an earlier flowering time, along with an earlier harvesting time, is in farm management: providing the opportunity for portfolios of varieties with different developmental timings so that the demands on labor and machinery can be better spread out during the season.
This advantage depends on the size of the farm with larger farms being able to grow more varieties and take further advantage of this strategy. However, the interest in dramatically altering phenology to widespread cultivation of wheat as a spring crop was generally low.
Drought was clearly acknowledged as a problem that is already occurring and is likely to persist in the future. With forecasted effects of climate change leading to increasing temperature and extreme weather events, there is going to be more risk in general associated with arable crop production. Therefore, the more options there are for flexibility and resilience, such as a wide range of wheat varieties with beneficial agronomic properties, the better.
The issue of taking a yield penalty in early flowering or early harvesting winter wheat does not seem to be an insurmountable obstacle in terms of whether farmers will take up a new variety. One interviewee suggested that a 5%-6% yield penalty could be acceptable if the variety opened up other opportunities, for instance, if the variety had a higher value for output such as a milling wheat.
Furthermore, it was recognized that there is a trade-off between possible penalty in yield from growing early flowering varieties versus loss in end-use quality or yield in later flowering varieties due to drought stress (or due to not being able to harvest in time). Yield must ultimately be balanced with the costs associated with its production (Swarbreck et al., 2019) so there is potential scope to introduce more balanced assessment and decision-making criteria. However, there is still the hurdle of having to obtain a yield within 2% of existing RL varieties in order to be included on the RL. Considering that 94% of varieties sold in the UK are on the RL at time of purchase (NIAB TAG, 2020), this is clearly a bottleneck for varieties that do not make the yield threshold but offer other favorable agronomic or production characteristics, such as early flowering.

| CON CLUS ION
In UK winter wheat varieties, alleles of major phenology genes are predominantly fixed (Bentley et al., 2014) which constricts the variation available due to major effects. Despite this, there is significant variation for flowering time in the wider northern European gene pool (Langer et al., 2014) which suggests that our defined 5-to 10-day earliness window is tractable to achieve using available genetic resources. We report that earlier flowering times are desirable for UK wheat producers particularly giving greater flexibility in more fluctuating seasonal conditions, mitigating drought stress and contributing to improving the logistics of crop management. It is therefore urgent to prioritize further research to identify additional genetic controllers of flowering response, plasticity of effects and the underlying genetics and physiology linking flowering, maturity and yield. The creation and field testing of near-isogenic lines in elite UK germplasm for genetic variation in phenology provides an ideal mechanism for geneticists and agronomists to work together to understand flowering time effects. This has been previously demonstrated to be effective for understanding the production impacts of developmental variation in wheat (Hunt et al., 2019). This will support future breeding strategies to deliver optimized flowering time to meet the evolving requirements of UK wheat production.

ACK N OWLED G M ENTS
We gratefully acknowledge the contribution of the NIAB regional, NIAB TAG and Velcourt Ltd. agronomists who gave their time to be interviewed for this study, as well as Stuart Knight, Clare Leaman, and Stéphanie Swarbreck at NIAB, Matthew Cobald at Velcourt Ltd. and Mark Dodds at KWS UK Ltd. for their valuable input and discussions. We also thank the editor and two reviewers for providing constructive improvements to the manuscript. We acknowledge support to HS from the Biotechnology and Biological Sciences

Research Council (BBSRC) Flexible Talent Mobility Account Postdoc
Placement grant scheme awarded to the University of Cambridge.
AB is supported by the BBSRC Cross-Institute Strategic Programme "Designing Future Wheat" BB/P016855/1.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
HS and AB designed the study. HS conducted the agronomist interviews and analyzed the responses. HS and AB wrote the paper.