Climate change has now become a reality and is affecting all aspects of life on Earth (IPCC, 2023). Its impact on agriculture is of particular concern, as humans rely on it for food production and it is also critical for the sustainability of many regions worldwide (Smith et al., 2014). While climate changes (such as increasing temperature, higher risks of drought, extreme events and combinations of deleterious environmental conditions) and their uncertainties have been predicted at the world level (IPCC, 2023), their impacts on the agricultural sector, involving complex interactions between biological, socioeconomic and political aspects, are difficult to assess and predict (Ruane & Rosenzweig, 2018). Similar to other crops, the suitability of growing wine grapes may evolve significantly across different regions. Given that the profitability of viticulture relies majorly on the link between the quality of wines and the pedo-climatic conditions of grape growing sites, the socioeconomic impact of climate change could be even higher for wine grapes than for other crops. In their review, van Leeuwen et al. (2024) report that most traditional areas of the coastal parts and lowland regions of southern Europe and Southern California may lose their suitability for producing high quality wines with economically sustainable yields. Meanwhile, other regions in the north of Europe and at higher latitudes in North and South America may become more suitable for profitable wine production. Globally, the increases in suitability are mainly related to the rise in average temperature, while the loss of suitability is linked to increased drought risks combined with extreme temperature events. In most traditional winegrowing areas of Europe, grapevines are predominantly grown rain-fed in environments with poor soils, which is considered a highly sustainable way of economical land development. In these locations, poor access to water resources limits the possibility of irrigation, increasing the vulnerability of grapevine to other biotic and abiotic stresses in severe drought. In order to maintain the suitability of these regions for winegrowing, understanding the physiological responses of grapevine to drought and the mechanisms of adaptation/tolerance should be a top priority.

In response to stress, plants have developed complex processes which support reproduction or survival and are controlled by sophisticated molecular regulatory mechanisms which can be specific or generalist. The response of a plant to a combination of stresses can be unique, and it can be much more complex than a response to a single stress (Ollat et al., 2023; Martínez-Lüscher et al., 2024). Simonneau et al. (2017), Gambetta et al. (2020) and Gambetta et al. (2024) described the diversity of the responses of grapevine to drought depending on the stage, duration and intensity of stress application, thus corroborating the assumption that adaptation can rarely be linked to a single type of process. As stated previously (Ollat et al., 2019; Ollat et al., 2023), the adaptation of a crop such as grapevine can be defined as the ability to maintain an optimal trade-off between yield, berry composition and the longevity of the perennial structures. From a general point of view, adaptation means either a process across generations leading to a new combination of favourable alleles or a status “to be adapted” which corresponds to a given combination of alleles (Cooper & Hammer, 1996). The resulting phenotype is constitutively adapted to, or changes with, environmental pressure. This ability to change is defined as plasticity (Bradshaw, 1965), which does not always result in a more adapted phenotype. Acclimation is defined as the ability of a plant to withstand abiotic and biotic stresses through modifications (Mickelbart et al., 2015). As in other species, there are numerous studies on the identification of genes involved in the responses of grapevine to abiotic stresses through transcriptomic approaches, corroborating the assumption that the mechanisms involved are complex (as reviewed in Delrot et al., 2020; Gomès et al., 2021; Martínez-Lüscher et al., 2024). However, only very few genes identified in grapevine, such as VaHsfC1 (Jiao et al., 2022), VvGST40 (Nerva et al., 2022), and VvEPLF9.1 (Clemens et al., 2022) have been functionally validated (as reported by Ollat et al., 2023), and their effects on whole plant performance in relation to adaptation have not been clearly demonstrated thus far.

To address these different levels of complexity and to apply them to crop improvement, Munns and Miller (2023) suggest that plant responses to abiotic stresses should be analysed in terms of capacities, thus providing a simplified framework for understanding successful adaptation strategies. When considering an abiotic stress as an external force applied to an object (in our case a plant), and not as the response itself, the authors defined a constraint as the effect of the stress on apparent change in volume, shape or length of the object, depending on its capacity to respond; for example, the effect of applying a certain amount of pressure (stress) to a wooden ball by squeezing it in your hand will not be the same as for a rubber ball. Like a wooden ball, adapted plants show little evidence of constraint or perception of stress; however, they adopt a strategy to manage the consequences, such as changing their shape or composition. Conversely, like a rubber ball, non-adapted plants respond directly to the constraint; for example, they may grow very slowly and barely survive in response to the same pressure or stress. By analysing the responses of the plants in terms of capacities, it may be possible to determine which processes need to be targeted in order to limit or overcome the impact of stresses. This may help prioritise traits and underlying genes for efficient plant improvement. Munns and Miller (2023) defined seven capacities which can be divided into three categories (Table 1) and showed how they are relevant to several stress responses.

As reported by Munns and Miller (2023), plants not only respond to drought by dropping leaves and thus decreasing their leaf area (asset management), but they also modify their leaf shape or inclination and their root-to-shoot ratio (shape shifting). While photosynthesis is reduced, affecting growth, carbohydrate pools can still be maintained, thus providing energy for cell division and maintenance (supply chain and energy production). Osmotic adjustment occurs, as does the repair of oxidised macromolecules (supply chain and repair). The regulation of water uptake (together with mineral nutrients) and water losses, and its efficiency in terms of biomass production are also fundamental (supply chain and exclusion). Interactions with soil microorganisms, such as arbuscular mycorrhizal fungi (AMF), contribute to these capacities. Long distance signalling regulates stomatal opening (also regulated at leaf level by short distance signals) and ensures a balance between shoot and root growth (communication), roots and shoots both playing a major role in the adjustment of plant water status (uptake versus transpiration). This may involve various strategies of adaptation, such as saving water versus increasing the explored volume of soil for water resources.

Moving from a conceptual framework to a real case, Gambetta et al. (2020) and Gambetta et al. (2024) described the complex responses of grapevine to drought. They highlighted the traits or processes that may be the most relevant for identifying adapted cultivars, taking into consideration the fact that relevant traits or processes can vary depending on the intensity and timing of the water deficit (Figure 1). Under increasing water deficit, the capacity of grapevines to grow and produce fruit (supply chain, asset management and shape shifting) is progressively affected. When water deficit is moderate, the effects on berry quality can be positive, in particular in red cultivars grown to produce red wines. Under severe levels of water deficit, the capacity to control gas exchanges through hydraulic and biochemical signals becomes key (communication, supply chain and asset management). Meanwhile, under extreme drought conditions, there is a risk of mortality (asset management, repair and exclusion). At any level of water availability, above and belowground plant traits will contribute to the responses of the whole plant.

Figure 1. Numerous traits involved in the responses of grapevine to increasing intensity of water deficit (Gambetta et al., 2024).

When water availability is unlimited, canopy area, maximal transpiration, stomatal conductance and root system size are the main drivers of water uptake (Lobet et al., 2014; Gambetta et al., 2020). When water becomes scarce, another set of relevant traits is required in order to regulate water losses through stomatal regulation and to maintain water uptake from the soil via appropriate root development and hydraulic architecture. Photosynthetic machinery appears to be very tolerant to mild water deficit, and varieties differ in their strategy to cope with photoinhibition (Medrano et al., 2003; Flexas et al., 2004). The accumulation of different kinds of osmo-protectants, in particular in leaves, allows turgor pressure to be maintained as plant water potential decreases (Gambetta et al., 2020). Finally, when drought becomes severe, the capacity to maintain the hydraulic conductance across the soil-plant-atmosphere continuum appears to be essential. Grapevines are hydraulically segmented and appear to operate most of the time within a ‘safe’ margin of water potentials (Charrier et al., 2018; Lamarque et al., 2023). Stem cavitation is extremely rare, while basal leaves and petioles are more vulnerable (Schultz, 2003; Tombesi et al., 2014; Gambetta et al., 2020). Basal leaf drop is a prominent symptom of drought in grapevine (Figure 2), and the variations in sensitivity of petiole or leaf conductivity under water stress has been reported to be intraspecific across Vitis vinifera (Tombesi et al., 2014; Martorell et al., 2015; Dayer et al., 2020). The perennial features of grapevines and the impacts of pluriannual drought are nevertheless rarely studied (Shtein et al., 2021). Cavitation fatigue (i.e., a decrease in resistance to cavitation after multiple cycles of cavitation and refilling; Hacke et al., 2001; Gambetta et al., 2020) should be considered, as well as season-to-season carry-over effects on bud fertility and stress memory mechanisms. If droughts should occur repeatedly, the resulting effects of a decline in carbon assimilation (lower gas exchange) and storage reserve replenishment – especially in roots – on vineyard lifespan should be investigated.

Figure 2. Leaf shedding symptoms of Tempranillo vines under drought situation in Cariñena (Spain), © Laure de Rességuier (Bordeaux Sciences Agro).

As a result of their general analysis, Gambetta et al. (2020) were able to define drought tolerance as a combination of four traits: maximal transpiration rate (E_max), stomatal regulation expressed as the Gs/Ψ_leaf curve, turgor loss point (Ψ_TLP) and the volume of soil explored by the root system. This combination defines a “grapevine’s stress distance” and could quantify (with a unit of time) how long a vine can function without water, under given environmental conditions, before reaching its critical water potential threshold.

Ongoing research aims to identify the most relevant traits and those that are the easiest to characterise, as well as to explore the variability that exists across available genetic resources, the physiological and genetic control mechanisms, and how the performance of ideotypes under future climatic conditions can be simulated. Some examples of this research is described in the following sections.

Dayer et al. (2022) characterised several hydraulic traits (E_max and plant stomatal conductance G_Smax, stomatal closure Gs90, extent of osmotic adjustment Ψ_TLP, leaf vulnerability to embolism (P50 leaf), the hydraulic safety margin (i.e., the difference between stomatal closure and 50 % loss of conductivity (PLC) due to embolism)) across nine genotypes (five Vitis vinifera cultivars and four Vitis accessions). These authors showed that all nine genotypes closed their stomata before leaf embolism development (HSM_P50 > 0), and that the variability of all the considered traits was greater among V. vinifera cultivars than among Vitis accessions. Using the soil-plant water transport model SurEau (Cochard et al., 2021) and based on the previously evaluated traits, the time taken for a mature grapevine to reach full leaf conductivity loss without water was calculated. It was shown that Grenache and Vidadillo (a minor Spanish variety) could last up to 150 days, followed by the Vitis labrusca accession (125 days). The shortest time (75-80 days) was estimated for Vitis candicans and V. vinifera cv. Yiannoudi (a Cypriot variety). Using a simulated library comprising more than 50,000 random trait combinations (eight hydraulic traits were considered: Gc_max, Gc₉₀, G_night, G_min, π₀, leaf modulus of elasticity, P50_leaf and PLC_slope), ideotypes that performed better than Grenache were identified (2.2 % of the simulations), and the 200 best-performing combinations (0.3 %) were called “Elites”. These Elite ideotypes were characterised by an average time to reach 100 % PLC of 173 days, linked mainly to lower Gc_max and G_min, a less negative Gc₉₀, a more negative P50_leaf, and a larger HSM_P12. Elite status was not associated to any specific single trait, confirming that drought tolerance estimated with the “grapevine’s stress distance” parameter depends on a trait syndrome (i.e., a combination of traits). When including belowground traits, such as rooting volume, in the modelling approach, it was possible to identify Super Elite ideotypes, which were characterised by an average time of 234 days for 100 % PLC. Finally, when the performance of the current Elite and Super Elite ideotypes were simulated under the most pessimistic future climatic scenario up to the end of the 21st century for six currently existing wines regions (that do not use irrigation), all the genotypes were projected to perform well in Bordeaux and Champagne up to 2070. In these regions, Grenache and Vidadillo are predicted to be unaffected until the end of the century, but to experience some leaf conductivity loss after 2050 in drier and warmer regions. Elites and some Super Elites may suffer only after 2070, especially in very hot and dry regions, such as the Napa Valley and Paso Robles in California. This study shows that, based on drought tolerant trait syndromes that prevent hydraulic failure, significant genetic improvement in terms of drought tolerance can be expected in grapevine, even within intra V. vinifera crosses. Other traits not included in this work, such as components related to carbon assimilation and root system characteristics, should also be investigated.

Despite processes at the soil-root interface being central to the control of plant water status (Javaux et al., 2021), belowground studies are rare, in particular in grapevine. Delval et al. (2024) showed that transpiration is controlled by a decrease in belowground hydraulic conductance (i.e., soil and root hydraulics) rather than by xylem cavitation in the trunk. Roots are a key component of plant adaptation to drought, in which a combination of structural and hydraulic properties can act across different time and spatial scales, and which is related to a whole spectrum of previously described capacities (Lynch, 2022; Ollat et al., 2016; Fichtl et al., 2023; Bernardo et al., 2025). According to Bernardo et al. (2025), the root system is still considered as the ‘dark side’ of a perennial plant: its structure and physiology have been little investigated, and identifying the combination of root traits that could maximise drought tolerance remains a significant challenge. Flor et al. (2025) have suggested that several rootstock properties related to xylem anatomy, hydraulic segmentation sensitivity, root growth and the control of the minimal conductance of the scion may interact to affect whole grapevine drought tolerance. This multi-level complexity of root systems (structural, architectural and functional, as well as the plasticity of these properties) and their inaccessibility limit our understanding of the most relevant traits associated with drought tolerance. Nevertheless, with modern phenotyping techniques and modelling approaches, advances are being made in the field.

Most adaptive responses are genetically regulated, meaning that breeding new cultivars for both scions and rootstocks is a powerful tool for adaptation to drought in the context of climate change; however, the contribution of epigenetic processes to mitigating the environmental effects must also be considered, especially in the case of a clonally propagated plant such as grapevine (Fortes & Gallusci, 2017; Berger et al., 2023).

Studies on the epigenetic mechanisms involved in grapevine development and stress response have been reviewed by Berger et al. (2023) and Venios et al. (2024). Among others, epigenetic regulation plays a role in stress memory and in the priming process. Plant priming describes the capacity of plants to modify their responses to biotic and abiotic stresses after initial exposure to mild stress or to eliciting molecules (Gallusci et al., 2023). In grapevine, stress memory has not been clearly demonstrated thus far. Nevertheless, Marfil et al. (2019) have observed that a combination of drought and UV-B during one season affects the epigenetic landscape over at least two seasons. In addition, it appears that vegetative progenies of grapevines initially grown in the field with different levels of water availability retain the ability to respond to water deficit depending on the phenotype of their mother plant (Paggay et al., 2022; de Deus et al., 2023). When exposed to a combination of heat and drought, primed plants showed changes in expression of the genes linked to epigenetic modifications, even after stress removal. Transcriptional responses have also been found to be modified the year after the application of the stress, with a general increase in DNA methylation for primed plants; however, the epigenetic control of memory seemed to be preferentially related to post-transcriptional regulation and histone modifications (Tan et al., 2024). First year analyses performed on Cabernet-Sauvignon cuttings showed that priming affects the survival capacities of plants regardless of plant size. There is now an urgent need for studying the consequences of priming on the intensity of stress symptoms across two or more stress cycles at metabolomic, transcriptomic and epigenetic levels.

This overview explores plant adaptation to drought from soil to atmosphere. Grapevine has been shown to develop many different strategies – especially as a grafted plant – for coping with a threat of future decreasing water availability. Adaptation to even a single stress event depends on many plant-related factors, such as their maximal transpiration and root hydraulic components. Thus, it is critical for scientists and growers to consider the plant as a whole, and across its entire lifespan. Different syndromes of traits could also be better adapted to specific drought scenarios and stress memory should been taken into account. The analysis of agricultural systems using a conceptual framework is highly recommended to help simplify the underlying complexity of responses and adaptive capacities. Modelling is potentially useful as an integrative tool to address this complexity and integrate a wide combination of traits to anticipate responses to future climatic conditions. Other integrative approaches can be implemented to link genotype to phenotype, taking into account epigenetic regulation and extended genotype, including interactions with microorganisms. Regulation hubs or the groups of processes related to whole plant hydraulicity, which contribute to building up the plant’s capacity to produce and survive, should be taken into account when identifying key traits and genes for the breeding of more adapted genotypes of both scions and rootstocks. In addition, it is important to take into account trait plasticity in various water availability conditions that are influenced by soil properties, climatic hazards and growing practices when studying the interactions between genotypes, environment and practices. Together, these approaches could help to provide systemic solutions to drought that leverage both breeding and management practices.

The authors acknowledge all the staff members of UMR EGFV, and the Ph-D students and post-doctoral fellows who participated to the cited studies. They also thank the funding bodies which support these multidisciplinary analyses.