The negative effects that climate change has on agriculture constitute a challenge, with the risk of extreme events being one of the key aspects of current climate research (Choudhary, 2015; FAO, 2016; IPCC, 2012; IPCC, 2018). The impact of climate change on viticulture has been extensively studied (Bernetti et al., 2012; Fraga et al., 2012; Mozell and Thach, 2014; Yzarra et al., 2015), with studies specifically targeting viticultural suitability (Hannah et al., 2013), grape and wine quality and production (de Orduña, 2010), irrigation strategies (Chaves et al., 2010), tillage treatments (Fraga et al., 2012) and grapevine growth stages in the face of changing climate (de Cortázar-Atauri et al., 2017; Greer, 2013; Jones and Davis, 2000). Wine production regions worldwide have experienced changes in climate structure, resulting in shifts of the timing of phenological events in the grape varieties cultivated, changes in grape chemistry and wine quality, and increases in the incidence of insect-borne diseases and grape-ripening disorders, such as berry shrivel (Krasnow et al., 2009; White et al., 2006; Zufferey et al., 2015a). According to grape producers cold and wet growing seasons, extreme heat conditions, rain during bloom and harvest delays are the most severe climate-related risks (Belliveau et al., 2006).

Scientists have shown that the current climate changes have altered the frequency, intensity, spatial extent, duration and timing of extreme climate events (IPCC, 2012; IPCC, 2018). An extreme climate event is defined as ‘the occurrence of a value of a climate variable above (or below) a threshold value near the upper (or lower) end of the range of observed values of the variable' (IPCC, 2012). There is some evidence that heat waves will occur more frequently (Meehl and Tebaldi, 2004) and drought will intensify in the 21st century during some seasons (mainly during summer) and in certain areas across Europe (Hov et al., 2013). Extreme events, such as late spring frost (Leolini et al., 2018) or extreme temperature and water stress, may significantly affect plant development (Gray and Brady, 2016; Hatfield and Prueger, 2015).

In cool regions, such as Switzerland, low temperatures often limit leaf and canopy photosynthesis and sugar production, although growth and sink activity of the fruit decrease to a greater extent than photosynthesis under low temperatures (Körner, 2003). Temperature and rainfall conditions are crucial before flowering. Low temperatures (< 15 °C) can lead to poor fruit set due to excess flower abortion. Extreme frost and rainy events during the flowering period can lead to substantial yield losses (Keller and Koblet, 1995). Following fruit set (post flowering development), the rate and duration of cell division in the berry pericarp are controlled by the number of seeds in the berry, as well as by climatic conditions. Extreme temperatures (very cold or very warm) may inhibit cell division and expansion. Cell division is mostly under genetic control, whereas cell expansion is predominantly driven by environmental factors (Keller, 2015). Extreme heat events combined with water stress sharply decrease cell expansion and yield. Nonetheless, leaf photosynthesis can adapt to the prevailing temperature at a given time during the season, particularly after flowering, and this so-called modulative temperature adaptation (Larcher, 1995) may occur within a few days or, sometimes, hours. Possible modifications of substrate concentrations or RuBisCO activity and structural alterations of the bio-membranes may explain differences in adaptability to increasing temperatures across cultivars. For grapevines, local acclimation to the prevailing temperature conditions can mitigate the effects of extreme heat events during the ripening period (Zufferey et al., 2000). However, a rise in temperature is often accompanied by an increase in canopy evapotranspiration, ultimately increasing the risk of water stress and physiological disorders, such as embolism events (e.g., disruption of the hydraulic conductivity of the vessels) (Zufferey et al., 2011), which may further inhibit plant photosynthesis and growth and reduce productivity (Dayer et al., 2017).

To avoid such consequences and maintain the yield and quality of vineyard harvest, adaptation strategies (Mosedale et al., 2016; van Leeuwen et al., 2019a) are needed. Without these strategies, extreme heat is expected to have detrimental effects on vine physiology and yield, even though some varieties are more tolerant of extreme temperatures than others (Santos et al., 2020).

In the Swiss Rhone catchment (Valais), permanent crop cultivation (orchards and vineyards) and livestock production are the most important agro-economic activities. Under the predicted climate change scenarios, the adverse effects of extreme heat events on Swiss vineyards are expected to become a threat in the upcoming decades (Fuhrer et al., 2014). Although there are some data on the increasing risk of spring frost damage in grapevines due to climate change in the Swiss Rhone (Meier et al., 2018), there remains a knowledge gap regarding the risk of extreme temperature events in grapevine plantations in this region.

To this end, the aim of this study was to address the following questions: (1) can the effects of extreme temperature events on grapevine phenophases be captured? (2) which phenophases are the most sensitive to extreme temperature events? and (3) which grape varieties are the most robust (or least sensitive) under extreme temperatures? The observed patterns may offer novel insights to aid wine producers in decision-making on vineyard management in the face of climate change.

The results in Figure 2 show that, on average, flowering takes place earlier and earlier. When comparing the respective temperatures of a given year before the flowering event, it is possible to observe an increase in temperature before the flowering event over the years (see, for example, Figure 3); this is obviously negatively correlated with the day of the year in which the phenological events took place; which is not surprising and is consistent with many studies. Climate-induced phenological shifts have been observed across Switzerland (Bigler and Bugmann, 2018). Scientists have shown that the current climate changes have led to changes in the frequency, intensity, spatial extent, duration and timing of extreme climate events throughout the world (IPCC, 2012; IPCC, 2018). Data from Austria, France (Chuine et al., 2004; Duchêne et al., 2010; Maurer et al., 2011), Germany (Menzel, 2005) and the Swiss Alps (Büntgen et al., 2006) have confirmed that the year 2003 was an unprecedented extreme year (García-Herrera et al., 2010). Many parts of Europe experienced record-breaking temperatures during July 2006, exceeding the values recorded in 2003 (Dankers and Hiederer, 2008). Furthermore, the winter season of 2006–2007 was estimated to be the warmest in the previous 500 years (Luterbacher et al., 2007). These extreme years are also confirmed by meteorolgical data set from Sion (SIO) and Evionnaz (both in Switzerland) that we used for this study.

Generally, temperature is a key factor affecting plant development (Went, 1953). The ultimate impact of temperature stress on yield or reproductive fitness depends on the developmental stage at which the high temperature stress occurs (Gray and Brady, 2016; Hatfield and Prueger, 2015), as well as on the variety (Martínez-Lüscher et al., 2016). Water stress also increases the vine's susceptibility to heat stress and drought during the bloom period is especially detrimental to fruit set, if it coincides with a heat period (Srinivasan and Mullins, 1981). The upper temperature limit for maximum yield formation in grapevines seems to be at 35 °C (Keller, 2015). Higher temperatures (> 35 °C) can produce a so-called heat-shock and protein deformation in berries with sunburn symptoms as an expression of oxidative damage from a combination of high light intensity and high temperature (Iba, 2002). The acclimation to low temperatures (< 15°C) - known as chilling acclimation - during the season leads to higher leaf photosynthetic rates and water use efficiency at lower temperatures, because of a downward shift in the optimum temperature for photosynthesis (Zufferey et al., 2000). Reproductive development is more vulnerable to chilling stress during the pre-flowering stage (Keller, 2015), because fruit sink activity is more sensitive to temperature and carbon shortage at this time than shoot sink activity. Previous studies mainly showed correlations between climatic variables (such as temperature) and flowering times, mostly by linear correlation or liner regression (e.g., Menzel, 2005) or model non-linear responses to temperature (e.g., de Cortázar-Atauri et al., 2010). An exception is the study from Siegmund et al. (2016), who first used appropriate techniques to identify periods prior to the growing season, where extreme temperatures events are statistically related to extreme flowering dates. In this study, we carried out an event coincident analysis (modified from Siegmund et al., 2016) to identify the way in which extreme temperature events control the timing of grapevine phenophases. Furthermore, we applied two other methods (quantile and robust covariance estimation) to compare the results of the modified Siegmund method and prove its novelty.

Using the phenological and temperature datasets, we identified the years when extreme events in both datasets coincided with one another. We showed that most of these coincidental events occurred after 2000 (mainly in 2007, 2011, 2014 and 2017; see Figures 4 and 7), which is the warmest decade on record since the beginning of the modern meteorological records (WMO, 2013; Arguez et al., 2020).

The analysis and results confirmed that our modified Siegmund method is more sophisticated than the other two methods, bringing added value to the detection of extreme coincidence events, which provides us new knowledge about the extreme events in the studied Swiss vineyard. By using this method, we discovered that bud development and flowering experienced more extreme coincidence events than the other phenophases and that bud development has significantly more extreme coincidence events than inflorescence merge and ripening of berries. In particular, the much larger influence of temperature on bud development compared to other phenophases has also been shown by other studies, even if they focused on other types of wine from other regions (Malheiro et al., 2013). Lethal freeze injury to buds is decisive during winter and spring in terms of potential yield. Low temperatures lead to cold acclimation during winter. Nevertheless, warm episodes during the acclimation period induce rapid de-acclimation, causing problems when a freezing event follows a mild winter, for example. Osmotic adjustment in buds, especially the adjustment of osmotically active sugars (sucrose, glucose and fructose), is essential for good acclimation, because it reduces cell dehydration and inhibits the nucleation and growth of ice crystals in bud cells (Keller, 2015). Deficient C- and N-reserves in permanent grapevine organs and a sudden increase in temperature in early spring can lead to early budbreak, increasing the risk of bud freeze and loss of fruit formation (Zufferey et al., 2015b).

We did not find any significant differences in the sensitivity of the tested varieties based on the occurrence rate of extreme coincidence events. Thus, we could only rank them based on the total number of extreme coincidence events they experienced during 1978–2018. Disregarding the two wine varieties Arvine and Gamay, from which we had less data, Chasselas and Chardonnay experienced the most extreme events (Robinson et al., 2012; Töpfer et al., 2009); these varieties can therefore be ranked as the most sensitive of the studied ones, which is in line with conclusions drawn by Robinson et al. (2012) and Töpfer et al. (2009). Zufferey et al. (2000) reported that photosynthetic rate was observed to increase with increasing temperature for Chasselas, compared with other European cultivars such as Riesling. Even when modulative temperature adaptation occurred (Larcher, 1995), with increasing water scarcity and extreme temperatures hydraulic failure was observed in the leaf and petiole vessels (xylem tissues) of Chasselas (Zufferey et al., 2011). It is also known that for Pinot noir, the climatic niche is narrow, and the average growing-season temperature for this variety is relatively low (Spring et al., 2010). Early-ripening varieties such as Chasselas, Pinot noir and Chardonnay could become more widespread globally, if they were to be cultivated in new, more poleward regions (such as Canada, northern Europe and Tasmania; Morales-Castilla et al., 2020).

The time series for Arvine and Gamay were too short, and when all the wine varieties were broken down to this short time, our results were not significant. From literature, it is known that Gamay is grown in cool-climate regions such as Canada and Switzerland (Robinson et al., 2012). While the white varieties (Chardonnay, Chasselas and Arvine) are poorly suited to the dry conditions (Zufferey et al. 2020), the cultivar Arvine is particularly sensitive to temperature increase and water scarcity events, thus influencing its aromatic compounds. When heat-susceptible cultivars, such as Arvine, are grown in warm climates or in hot growing seasons with intense solar radiation, they often develop poor aromatic compounds in berries and have a negative impact on wine quality (Zufferey et al., 2020). More data on Gamay and Arvine would be necessary to prove these statements using our methodology. Syrah (and Pinot noir) were the least sensitive of all the wine varieties examined in this study (see Table 3 and Figure 7). Late-ripening varieties, such as Syrah, Grenache and Mourvedre, are projected to become much more widespread in current global winegrowing regions if temperatures rise by 2 °C (Morales-Castilla et al., 2020).

Several attempts have been made to evaluate drought tolerance and to rank grape varieties (Parker et al., 2011; Parker et al., 2013), in order to help the variety best adapted to climate change. However, these data are not yet available for all . Wine growers could therefore maintain their income by replacing the grapevines they cultivate with other varieties or crops (Duchêne, 2016; Duchêne et al., 2014).

Although grapevines have several survival strategies (e.g., deep root systems or efficient stomatal control), viticulture is strongly dependent on climate (Fraga et al., 2012). With more extreme events being expected in Europe (Beniston et al., 2007; Hov et al., 2013; IPCC, 2012; IPCC, 2018), additional methods will be required in order to adapt to and mitigate climate change (Soja et al., 2011; van Leeuwen et al., 2019a; van Leeuwen et al., 2019b). Adaptation to higher temperatures include changing plant materials and modifying viticultural techniques. van Leeuwen et al. (2019a) have outlined the best available practices to make vineyards more resilient to drought, including planting drought resistant plant material (rootstocks), modifying the training system, or selecting soils with greater water-holding capacity.

The main purpose of the study was to capture the coincidences of extreme temperature and phenological events of grape varieties in Leytron, Switzerland. The comparison of various methods showed that the methodology of Siegmund et al. (2016) seems to be the most appropriate for studying extreme coincidences. The results showed that there were much more such events between 2003 and 2017 than in earlier years (1978–2018). Some phenophases (bud burst and flowering) are less robust in response to extreme temperature, but only significance results were obtained between bud development and inflorescence merge and ripening of berries. While we found differences, we did not find any statistically significant evidence of some grape varieties being more robust or sensitive to temperature extremes than other varieties. However, it might be possible to obtain statistically significant results when comparing grapevine varieties using the proposed methodology on larger data sets from other regions. The observed patterns in our study could provide new insights for winemakers to help make decisions about vineyard management in the face of climate change. However, in order to generalise the statements and findings, this research and methodology would need to be applied to other datasets from different grape varieties in other regions.