The study was developed in a commercial vineyard in an emerging wine region on the Atlantic side of southeastern Uruguay (Garzón, located in the Maldonado Department). The area's climate is classified as temperate, with temperate nights and moderated drought, corresponding to the ISA1IHA3IFA2, climatic group according to the “Multicriteria Climatic Classification” method for vineyards (Ferrer et al., 2007; adapted from Tonietto and Carbonneau (2004)). The grape ripening period in southeastern Uruguay is characterised by temperate-maximum temperatures, and the daily mean temperature during the growing season is 18.9 °C ( Fourment et al., 2020).

Four temperature sensors (Tinytag data loggers, Gemini, UK) were installed in plots of Albariño in contrasted topographical situations, which were classified into two classes, i.e., High (H) and Low (L) exposure to the Atlantic Ocean based on spatial climatic differences (Table 1 as mentioned in Tachini et al., 2023a for greater details) (Figure 1B,C). The four plots have the same type of soil, classified as Brunosol by the National Classification. This study shows results over five consecutive growing seasons from the 1st of September to the 1st of March of 2018–2019 to 2022–2023. This refers to the year of harvest, especially from 2019 for the 2018–2019 study season, up to 2023 for the final study season of 2022–2023.

Figure 1. Location of Uruguay in South America (A); location of the Rocha Weather Station from INUMET and the four plots at the different exposure to the Atlantic Ocean (High and Low exposition based on contrasting topographic situations) (B); zoom on the four plots (High and Low exposition) and the land’s topography (C).

Seasonal variability in temperatures and precipitation in the region was evaluated using daily data obtained from the Rocha weather station (Figure 1B), provided by the Meteorology National Institute (INUMET) network adhering to the World Meteorological Organization (WMO) standards. Regional climate variability was evaluated during the growing seasons over the five consecutive vintages (2019 to 2023). The climatic indicators used were precipitation during the growing season (PP), precipitation during ripening (January and February), number of days with precipitation (NPP), average growing season temperature ( Jones, 2006), thermal amplitude during summer (January and February) (THA), growing degree days from 1st September to 28th February (GDD) ( Winkler, 1974), cool night index as mean minimum temperature of February (CNI) and the numbers of days with temperatures above 35 °C (ND35).

Spatial variability of temperature within the vineyard was further assessed using hourly temperature data measured by temperature sensors (Tinytag dataloggers, Gemini, UK) through the calculation of bioclimatic indices for each season, at each plot, such as GDD; THA; CNI; season minimum temperature (TN); season maximum temperature (TX); number of days with temperatures above 30 °C (ND30); number of days with temperatures above 35 °C (ND35); maximum temperature in January (TXJan); minimum temperature in January (TNJan); maximum temperature in November (TXNov); minimum temperature in November (TNNov). To investigate the influence of sea breezes, we analysed the number of hours with temperature decreases of varying intensities (0.5 °C and 1.0 °C) during peak temperatures between 11:00 and 16:00 Local hour (LH) as mentioned in Fourment et al. (2014). Thus, the variables Freq.H-0.5 and Freq.H-1 represent the temperature drop of 0.5 °C or 1 °C during heat waves. We considered heat waves as a period of three or more consecutive days with a maximum temperature of 32.3 °C ( De Mello and Renom, 2024). Yearly climatic variables were used to study the grapevine's response to seasonal climate variability. The climatic variables obtained for each plot were used to assess the impact of spatial temperature variability over the vintage.

According to PlantGrape (2024), Albariño is a moderately vigorous variety with a horizontal growth habit that requires careful trellising. It is not highly productive and benefits from cane pruning. Well-suited to acidic, granite, and sandy soils, Albariño is relatively tolerant to Botrytis rot ( PlantGrape, 2024) and produces elegant wines with good acidity. Consequently, Albariño cultivation in Uruguay has increased by 24 % over the past five years, now comprising 1.6 % of the country’s vineyards ( Instituto Nacional De Vitivinicultura, 2024).

In this study, we analysed four plots of high-quality Albariño grapevines planted in 2009 (13 years old), which were trained to a vertical shoot positioned (VSP) training system and pruned using the double Guyot method. The planting density is 2 m between rows and 1 m between vines resulting in 5,000 vines per ha. Yield components were assessed for each plot using two replicates of seven contiguous vines. For each vine, we assessed the total number of clusters, yield per vine (kg/vine), yield per vine without Botrytis (kg/vine) (rot incidence), berry weight (mean value of the 250 samples of berries, the description is below), pruning weight (kg/vine) in winter, and Ravaz Index (Fruit weight/kg per pruning weight/kg) ( Ravaz, 1911).

During the ripening period, berry samples were collected weekly, as outlined by Carbonneau et al. (1991). Grape berry composition measures undertaken were sugar content (SC) (g/l) by refractometry, total acidity (TA) (g H ₂SO ₄/l) by titration, and pH by potentiometry, according to the O.I.V. protocol (1990). To analyse phenolic potential (Total Phenolic Index of grapes), the Glories and Agustin (1993) protocol was followed with a Shimadzu UV-1240 Mini (Shimadzu, Japan) spectrophotometer. These variables are important to assess fruit quality and winemaking potential ( González-Néves et al., 2010).

Univariate statistical analysis (ANOVA) was used to evaluate every climate and agronomic variable to determine the influence of season and exposure to the Atlantic Ocean at different confidence levels. We used two variation factors: plot exposure to the Atlantic Ocean named as plot site—PS (H and L exposure to the Atlantic Ocean) to assess local climate impact, year—Y (2019 to 2023 seasons) to assess seasonal impact; and we calculated the interaction PS × Y. The variance percentage was calculated for each variable between consecutive years. To analyse the plasticity of Albariño through the relationships between local climate and Albariño agronomic performances and oenological potential, a Pearson correlation matrix was made. Multivariate statistical analysis was performed to elucidate correlations between climate and agronomic data. Principal component analysis (PCA) was conducted to analyse the dataset to differentiate the plot's location and season effect. To demonstrate relationships between yield components and berry composition variables over the five seasons, plot location and season were analysed separately to identify the general variability effects. The statistical analysis was performed using Infostat® and OriginLab® programs.

The region's climate variability was strongly associated with precipitation patterns and temperature-high extremes (Table 1). On average, the growing seasons received 584 mm of rainfall over the five-year period. Seasons 2019 and 2022 received greater rainfall than the average during the season (+348 and +208 mm, respectively) and the ripening period (+92 and +128 mm, respectively). Only 2019 had a greater number of rain days. During the five seasons, the thermal indicators that showed statistical differences were THA, GDD and N35. 2020 and 2022 had the highest values in THA during ripening (1.09 and 0.59 °C more than the average, respectively), while in terms of heat accumulation 2023 was the season with higher GDD values (191 °C more than the average). On average, Rocha Weather Station showed 9 days with temperatures above 35 °C and 2.2 days in 2023.

The average values of the thermal indices calculated using data from the sensors installed in the vineyards showed a significant season effect, as well as a significant plot site effect for 8 of 13 indicators (Table 2).

Within the five growing seasons, 2023 was the hottest and driest, during which the greatest differences between plot locations were found (spatial variability). For example, under the warmest seasonal conditions of the study period, the impact of sea breeze was greater on plots exposed to the Atlantic Ocean, recording 6.5 hours more than others with a temperature drop at midday (Table 2). Other bioclimatic indices were influenced by ocean exposition: minimum temperatures, such as the Cool Night Index, minimum temperature during the growing season, January minimum temperature, and November minimum temperature (Table 2).

Albariño yields were, on average, 1.7 kg per vine, with similar berry weight between years (1.3 g) and similar cluster number (19.2 clusters per vine on average). The yields with and without Botrytis rot achieved in the plots were similar, showing a low rot incidence in this cultivar. Only in 2022 was Botrytis rot observed on grapes, resulting in yield losses (Table 3).

Differences in yield components between plot sites (High and Low exposure to the Atlantic Ocean, named H and L) were only observed in some seasons. Specifically, yield per vine was only statistically different in 2021, when H presented 0.72 kg per vine more than L, and cluster number in 2022, when L had 7.6 more clusters than H. Yield without rot was significantly different in 2021, when H plots had a higher rot incidence compared to L (0.72 kg). Regarding pruning weight, there was no consistent biomass production at the sites. In two seasons, 2020 and 2023, H plots had greater pruning weight than L (0.21 kg difference on average). Finally, vine balance (measured as RI) was statistically different in the last two seasons of the study (2022 and 2023), where L showed higher values compared to H (1.57 higher on average).

Table 4 shows the differences obtained in berry composition at harvest. Significant differences were found on average between plots in two variables only, for example, final sugar content (measured as °Brix) and total acidity; particularly in 2023, berries harvested in L had a sugar content of 0.9 °Brix higher than those harvested in H.

In terms of acidity, the 2020 and 2023 vintages exhibited greater differences between plots. Berries from H plots obtained higher acidity content than those from L plots (+0.98 g and +1.64 g in 2020 and 2023, respectively). Berry pH and total phenolic index showed no significant differences between years and plot locations.

The Pearson correlation matrix made to elucidate the relationships between local climate and Albariño agronomic performance and oenological potential showed that berry composition variables were more affected by exposure to the Atlantic Ocean than yield components variables. Berry composition of Albariño showed significant differences in plots distinguishing between High or Low Atlantic Ocean exposition (Figure 2). Berries with higher sugar content (°Brix levels) and lower total acidity were associated with L plots experiencing greater thermal amplitude during the ripening period and higher January minimum temperature ( r = 0.62; p = 0.003; and r = 0.83; p = 0.0001, respectively) (Figure 2A,C). pH and berry weight resulted in significant differences between plots experiencing extreme temperatures as maximal temperature during the growing season (Figure 2B) and number of days with temperatures above 30 ºC (Figure 2D) ( r = 0.75; p = 0.0001; and r = –0.67; p = 0.01, respectively).

Figure 2. Plasticity assessment between Albariño plots at different exposures to the Atlantic Ocean (High – dark grey circle; and L – light grey circle). (A) °Brix at harvest and Thermal Amplitude during summer (r = 0.62; p < 0.003), (B) pH at harvest and maximum temperature in the Growing Season (r = 0.75; p < 0.0001), (C) Total Acidity at harvest and January minimum temperature (r = 0.83; p < 0.0001), (D) Berry weight at harvest and Number of days above 30ºC (r = –0.67; p < 0.01).

PCA analysis of Albariño performance comparing plots, location and years showed that berry composition variables were more significant than yield components (70.7 % and 63.4 % of the total variability, respectively) (Figure 3). In terms of yield components, PCA showed that H plots were separated from L plots and were highly associated with cane weight, yield, number of shoots per plant and number of clusters (Figure 3A). When the season was considered, there was no clear association with yield measurements, except for the 2022 vintage where yield was associated with the incidence of bunch rot (Figure 3B).

A clear discrimination between berry acidity and sugar content was determined between plots with different expositions to the Atlantic Ocean (Figure 3C), where higher °Brix were reached in L plots. Also, we found discrimination between final berry composition and season (Figure 3D), showing an interesting finding through the five studied years: 2020 was associated with the total polyphenol index, and 2022 and 2023 were associated with total acidity.

Figure 3. Principal Component Analysis. (A) Yield components variables and plot location at different exposures to the Atlantic Ocean (High as H and Low as L), (B) Yield components variables and year, (C) Berry composition variables and plot location, (D) Berry composition variables and year. For the four figures, the points represent each plot.

Among the five growing seasons studied, 2019 and 2022 were classed as wet, with precipitation exceeding the average by 348 mm and 208 mm, respectively. For viticulture in Uruguay, having lower rainfall during ripening is desirable to mitigate the risk of disease. Despite the high water availability in 2019 and 2022, only 2022 experienced extreme precipitation during berry ripening (128 mm above average), resulting in Botrytis rot in Albariño. In the 32 years from 1991 to 2023, Fourment et al. (2024) observed no statistical trend in precipitation (increase or decrease) in eastern Uruguay. However, the decreasing trend in rainy days combined with an increasing trend in dry days, raises questions about the need to adapt grapevines to water deficit conditions, even in the presence of wet seasons. The last two seasons of the study, particularly the summer of 2023, were influenced by ENSO La Niña, leading to a severe drought compounded by a heat wave as experienced in other countries of central South America ( Arias et al., 2023). While the reduced amount of rainfall cannot be directly attributed to climate change, the higher regional temperatures can contribute to decreased water availability ( Arias et al., 2023). This suggests that climate change likely exacerbates agricultural drought by adding heat stress to water stress on crops ( Arias et al., 2023). Considering viticulture, such heat and water stress conditions can reduce berry yield, inhibit vine canopy development, and alter grapevine composition and typicity ( Ferrer et al., 2014; van Leeuwen and Destrac-Irvine, 2017), thereby impacting wine characteristics. This highlights the need for viticulturists, consultants, and oenologists in Uruguay to adapt to climate variability and extreme events to maintain vineyard productivity and sustainability, in which case, cultivar plasticity is crucial to a better adaptation to climate conditions.

Despite annual variability, the average precipitation during the studied years was sufficient to achieve vine water requirements ( Williams, 2014), except in 2023, the driest season of the past 32 years ( Tachini et al., 2023b). The 2023 season, being the driest and hottest, exhibited the greatest differences between plot locations, indicating significant spatial variability.

The vineyards' exposure to the Atlantic Ocean resulted in differences in thermal indicators between plots with higher and lower exposure. The moderating Atlantic Ocean’s influence is evident through various indices: heat accumulation (GDD), average temperatures during the growing season (TN, TX), specific phenological periods like flowering (TXNov and TNNov) and ripening (CNI, TXJan, TNJan), extreme temperatures (ND30 and ND35) and the thermal impact during potential development of sea breeze circulation (Freq.H-0.5 and Freq.H-1). For eastern Uruguay, the thermal amplitude is not a major mode of variability, unlike the southern wine region as demonstrated by Fourment et al. (2017). However, the CNI, TNNov and TNJan indicators displayed greater spatial variability, indicating that minimal temperature may be a crucial factor in distinguishing thermal conditions across this region over both space and time.

The sea breeze's influence on the spatial variability of temperature was significant in the region during the warm season of 2023, as shown in Table 2 in the frequency of the number of hours with a decrease in temperature at different intensities (0.5° and 1.0°) at the time of maximum temperature (between 11:00 and 16:00 Local hour), corroborating results over the South Africa vineyards ( Bonnardot et al., 2005). Fourment et al. (2014) showed that during berry ripening, a vineyard ~30 km inland from La Plata River spent less time (44 hours) experiencing a temperature decrease in the afternoon linked to sea breeze development, compared to Garzón vineyards (53 hours for H plots and 35 for L plots). This indicates greater differences within the region than in two different coastal regions. Garzón vineyards experience an average of 113 hours of temperature drop with greater exposure to the Atlantic Ocean, compared to 72.3 hours with less exposure. Although differences in ocean exposure do not significantly differentiate plots, the high thermal impact is statistically differentiated during the hottest summer. The small topographic differences between plots do not affect local events like sea breezes, but H plots are more impacted during dry seasons, showing greater temperature spatial variability. This finding is crucial for understanding climate change impacts, as vineyards exposed to sea breezes can experience reduced extreme temperatures during ripening, minimising thermal stress and preventing degradation of primary and secondary metabolites caused by high temperatures ( Carey et al., 2003; Sadras and Moran, 2012; Poni et al., 2018).

However, considering mean temperature conditions over the season (as expressed by GDD), cooler conditions are experienced in L plots closer to the ocean. Thus, the drop in temperature associated with the arrival of the sea breeze inland could be minimised compared to the data recorded in H plots. During heat waves, evaluating the duration that grapevines remain below critical thermal thresholds (such as 30 °C or 35 °C) can improve the precision of our assessments, particularly when considering the cooling effect of sea breezes. This aspect presents a potential bias in our methodology for studying the influence of sea breezes in the coastal vineyards of the region. To address this, further investigation is needed, potentially incorporating regional atmospheric simulations similar to those previously conducted in the vineyards of Southwestern Cape Province, South Africa ( Bonnardot et al., 2002; Bonnardot et al., 2005; Soltanzadeh et al., 2016).

Similarly, bioclimatic indices for viticulture in Uruguay should be further explored in relation to the ENSO phases, like the research conducted by Lemos Kaltbach et al. (2024) in Brazil. This approach will improve our understanding of the impacts of ENSO on Uruguayan viticulture and contribute to developing more resilient vineyard management practices.

To ensure sustainable production systems, a medium-term study of the adaptability of cultivars in consecutive growing seasons in each terroir is necessary. An analysis of five consecutive growing seasons of Albariño, a cultivar actively promoted in Uruguayan viticulture has demonstrated its potential due to its favourable agronomic and oenological performance in eastern Uruguay.

The Albariño cultivar in Eastern Uruguay has proven to be productive, averaging over 7 tons per hectare, and relatively stable over five seasons. Low temperatures during flowering (November in Uruguay) could impact fruit set and thus final yield in the season ( Keller et al., 2010) and cluster number of the next season ( May, 1961). This was the case in the spring of 2021, with November minimum temperatures being 0.9 °C and 1.1 °C lower than the five-year average in the H and L plots, respectively. These cool spring conditions significantly affected the final yield, with further spatial differences: L plots produced 0.72 kg per plant less than H plots. The lower maximum temperatures during spring 2021 could explain the limited cluster number of the 2022 season for both plot locations, with an average difference of 7.6 clusters per vine between H and L plots. In terms of berry quality, amongst the wet summers of 2019, 2021 and 2022 (receiving 283, 218 and 319 mm of precipitation, respectively, over January and February), Albariño experienced yield losses caused by Botrytis rot incidence in 2022 only, showing some tolerance to bunch rot. This result has relevance to climate change scenarios, as increased precipitation is projected during summer and autumn ( Intergovernmental Panel on Climate Change, 2023).

In terms of Albariño oenological potential for the five studied vintages, the average berry composition at harvest reached 202.7 g of sugars (20.8 °Brix) and 5.95 g H ₂SO ₄.l ^-1 of total acidity, which are ideal values for producing finely balanced wines. Sugar content exhibited significant statistical differences between years ( p < 0.001). As the viticulture sector has started to prioritise sustainability by implementing national production programs and certifications for sustainable grape management practices ( Instituto Nacional De Vitivinicultura, 2024), these findings are highly relevant. Albariño’s consistent yield levels and high-quality berry parameters under temperate and sometimes humid conditions suggest that it is a variety well-situated for continued planting in Uruguay. In the context of climate change, the moderate temperatures of the Atlantic eastern terroir, combined with the production of a white cultivar that can express style and quality, represent two complementary adaptive strategies for mitigating climate change impacts, as proposed by Naulleau et al. (2021).

Assessing cultivar plasticity under different plot climatic conditions can help viticulturists choose the best-suited cultivars to adapt to climate change and variability. This evaluation also helps improve disease tolerance, optimise vineyard management practices (such as pruning and harvest timing), and maintain berry and wine quality and typicity. Multivariate analysis of yield and berry composition data, considering plot exposure and seasonal climate, has revealed that: 1) yield components have a lower impact on vineyards exposure compared to the berry composition at harvest; 2) the correlations between berry composition and plot exposure are greater under warmer and drier conditions.

When yield components are analysed as a response to Albariño plasticity of local climate, our results show differences in yield (g/plant), number of clusters and cane weight of the vines from the plots of high exposure (Figure 3A), higher yield (0.8 kg higher on average) with lower clusters per vine (1 less cluster) and greater biomass (186 g more on average) (Table 3). However, the season effect is more complex because yield is not only driven by factors in the current season ( May, 1961), and the previous season temperature could affect yield components, as discussed above.

Albariño berry composition at harvest shows sensitivity to ocean exposure in two primary metabolites (sugar and acidity) which is amplified between plots under drier and warmer conditions such as 2023. These results are not in agreement with Wu et al. (2019), who argue that moderate temperature increases did not impact primary metabolites in Sauvignon blanc. A demonstration of Albariño plasticity is shown in Figure 2A,C, where the relationship between different temperature variables and berry composition at harvest is evident. Vineyards with lower ocean exposure exhibit higher sugar content in berries, a trait that is highly correlated with the thermal difference between day and night during ripening. This increased thermal amplitude, as analysed by Kliewer (1973), acts as a precursor for higher sugar content in berries due to more photosynthetic activity.

Conversely, Albariño vineyards with higher ocean exposure led to greater total acidity (TA) in berries, due to less extreme temperatures. Malic and tartaric acid are the two main organic acids in grape berries. Only malic acid is metabolised by the ripening berry, and thus, its content is impacted by temperature ( Rienth et al., 2021). The correlation between TA and minimum temperature in January highlights the impact of plot location on berry composition. While Wu et al. (2019) suggested that moderate temperature increases do not significantly influence sugar and acid metabolism in Sauvignon blanc cv., our findings indicate that meso-climate differences do affect these compounds in Albariño. As reported by Sweetman et al. (2014), the degradation of malic acid concentration is accelerated by high temperatures, and it depends on thermal amplitude.

Other variables, such as pH and berry weight, do not demonstrate clear plasticity to local climate conditions. The relative stability of pH across different seasons and sites emphasises its role in determining the optimal harvest time, which explains its consistent levels despite varying climatic conditions.

Fourment et al. (2017) reported similar results for the red cultivar Tannat in southern Uruguay, highlighting the influence of temperature spatial variability on secondary metabolite production. The same results were found in the Atlantic eastern region for the same red cultivar in Tachini et al. (2023a). One limitation of our present study is that we only measured the total phenolic index as a secondary compound. It would be beneficial to explore the impact of different temperature thresholds on aroma and aroma precursors in grapes, as suggested by Wu et al. (2019) and Carey et al. (2003). The variability observed in only two contrasting local climate conditions (H and L) needs to be further investigated. For instance, the indirect method of analysing two contrasting environments can be considered a localised phenomenon, as discussed by Bonada and Sadras (2014). However, the observed contrasts across growing seasons, due to interannual variability and differing levels of Atlantic Ocean exposure across locations, improve our understanding of this economically important variety for both the country and the region.