Abstract
The effects of moderate irrigation rates on vegetative growth, vine evapotranspiration, yield, and grape and wine composition were studied during six consecutive seasons in a mature vineyard planted with Vitis vinifera cv. Tempranillo in Requena, Spain. Vines were spur-pruned and trained to a bilateral cordon. Rain-fed vines received a yearly average rainfall of 368 mm, of which 169 mm occurred from April to harvest. Irrigated vines on average received 86 mm per year of additional water applications. Irrigation increased vegetative growth and vine evapotranspiration. As a result, yield was 31% higher in the irrigated vines. This increase in yield was primarily due to larger berry size and was correlated with vine evapotranspiration estimated by soil water balance. Irrigation did not alter the balance between the vine demand and the supply as indicated by the similar level of yield to pruning weight and leaf area to yield ratios observed in both irrigated and nonirrigated vines. On average over years irrigation had some minor negative effects on wine composition. It altered the balance between malic and tartaric acid, increasing the former and decreasing the latter. Irrigation also led to an increase in wine pH that together with a slight decrease in anthocyanin concentration reduced color intensity by 18%. However, the effects of irrigation on must and wine composition were largely different among years, probably because of the different rainfall amount and crop levels. Thus, under high crop level, irrigation tended to mitigate the negative effects of increasing yield on wine alcohol content.
Increasing soil water availability to plant by irrigation often increases crop biomass and yield (Vaux and Pruitt 1983). In grapevine, it is important to define the effect of irrigation on yield but particularly also on fruit and wine composition.
Generally speaking, irrigation is a common cultural practice in the viticulture of New World countries, while in Europe its use for wine production is still somewhat restricted or even prohibited based on a common, and often not scientifically proven, consideration that irrigation detrimentally affects wine composition. In Spain, for instance, irrigation of grapevines for wine production was forbidden by law until 1996, but in the last decade irrigation in vineyards has steeply increased. Yet in some areas, water applications after veraison are still prohibited.
Supplying irrigation to ensure the potential vine evapotranspiration normally reduces wine quality (Williams and Matthews 1990), perhaps because of an increase in berry size through irrigation. If other berry characteristics, such as skin thickness, are not affected by improving vine water status, then larger berries would have a lower skin to pulp ratio. That leads to a dilution of the main berry quality components that are localized in the skin. Water stress during the period from fruit set to veraison heavily reduces fruit size (McCarthy 1997) because of the detrimental effect of soil water deficit on early fruit growth (Ojeda et al. 2001) that cannot be recovered even if water supplies return at full dosage later in the season (Poni et al. 1994a). On the other hand, late season water restriction reduces fruit cell enlargement and water accumulation (Smart and Coombe 1983) and in general has a less detrimental impact on final berry size than early season water stress (McCarthy 1997).
Irrigation might also indirectly affect berry quality because of increased and prolonged vegetative growth. After veraison, shoot growth may compete for the carbohydrates available for fruit ripening. Increased vegetative growth might also impair cluster microclimate, particularly fruit light exposure (Smart et al. 1985). In other cases, irrigation has led to a delay in obtaining desirable sugar levels (Bravdo et al. 1984). However, reports also show that severe water stress might be detrimental to fruit quality because of poor canopy development and reduced leaf assimilation rate and thus an inadequate vine capacity to ripen the crop (Hardie and Considine 1976), particularly under high yield level (Freeman and Kliewer 1983).
Regulated deficit irrigation can be applied as a strategy to reduce the possible negative impact of irrigation on wine quality. In Europe, irrigation tends to be reduced after veraison, while in Australia deficit irrigation generally is applied during the period from fruit set to veraison (McCarthy et al. 2000). These different strategies may be due to different vineyard management styles and climatic conditions. Thus, with relatively high crop level, as in many Australian vineyards, reducing vine water status after veraison might be very detrimental to wine composition because of the high yield levels. In Europe, where traditional vineyards have a lower yield, a source limitation after veraison because of mild vine water stress should be less detrimental and might help to reduce berry size and its water accumulation. Because of the high dependence of fruit quality on various environmental and endogenous factors (Jackson and Lombard 1993), the overall effect of irrigation might change according to other cultural practices such as training system or crop level (Bravdo et al. 1984, Poni et al. 1994b).
The cultivar Tempranillo is originally from northern, cool regions of Spain and is at this time the most widely cultivated cultivar for production of red wines throughout Spain. It is reputedly sensitive to water stress and prone to early leaf senescence (Gómez del Campo et al. 2000). Previous research has shown some beneficial effect of irrigation on fruit ripening, mainly increased berry sugar concentration (Esteban et al. 1999), but elsewhere has shown a decrease in the concentration of skin anthocyanins (Esteban et al. 2001).
The main objective of this research was to test an irrigation regime based on moderate application rates, particularly reducing water application from veraison to the end of the season. We hypothesized that inducing a certain degree of vine water stress after veraison would increase yield but minimize the possible negative dilution effects of irrigation on grape quality. We expected that under high yield irrigation might have a positive effect on berry ripening and wine composition.
Irrigated and nonirrigated vines are compared in terms of vine evapotranspiration, water relations, vegetative growth, yield, and fruit and wine quality. The interaction between irrigation and crop level is also analyzed, taking advantage of year-to year variation on yield values.
Materials and Methods
Site description and experimental design.
The experiment was carried out during six consecutive seasons (2000 to 2005) in a Tempranillo (Vitis vinifera L.) vineyard planted in 1991 on 161-49 rootstock at a spacing of 2.45 by 2.45 m (1666 vines/ha). The vineyard is located near Requena (39°29′N; 1°13′W; elevation 750 m), Valencia, Spain. In 2000, a drip-irrigation system was installed and vines trained to a vertical trellis on a bilateral cordon system oriented north-south. Canopy-management practices, all manually performed, included shoot thinning and shoot-tip cutting. Shoot thinning was carried out each year according to vineyard manager goals, which led to a different number of shoots and hence different number of clusters collected among years. Leaf removal was not carried out, and cluster thinning to leave one cluster per shoot was only performed in 2005. All treatments were fertilized at a rate of 30-20-60-16 kg ha−1 of N, P, K, and Mg, respectively.
The soil at the site is a Typic Calciorthid, with a clay loam to light clay texture, highly calcareous, and of low fertility (0.66% organic matter, 0.04% nitrogen). The soil has a deep soil profile (>2 m), available water capacity is ~200 mm m−1 and bulk density 1.43 to 1.55 t m−3.
Budbreak for Tempranillo in the region usually occurs by mid-April and flowering by early June; veraison is reached by early August, with harvest during late September and leaf fall at the beginning of November. Climate is continental and semiarid with average annual rainfall of 430 mm, of which ~65% falls during the dormant period. Weather conditions were measured with an automated meteorological station located in the plot, and reference evapotranspiration (ETo) was calculated with hourly values by the Penman-Monteith formula (Allen et al. 1998).
Irrigation treatments.
Two treatments were carried out from 2000 to 2005: (1) rain-fed (nonirrigated) vines, which had not been irrigated since planting, and (2) irrigated vines. Each treatment had six replicates in a randomized complete block design. Each plot consisted of 10 rows with 9 vines per row and the surrounding perimeter vines used as buffers.
Irrigation normally started in June. The irrigation amount applied was based on fulfilling 100% of crop evapotranspiration, estimated as the product of reference evapotranspiration (ETo) and crop coefficient (Kc). The seasonal Kc employed varied with the phenological period and the expected pattern of leaf area development. Thus, from June to July, Kc gradually increased from 0.08 to 0.30. After veraison, the objective was to induce a moderate soil water deficit, therefore applied water amounts were 0.2 of ETo.
Drip irrigation was applied with two pressure-compensated emitters of 2.4 L h−1 located at 60 cm on each side of the vine. Irrigation was applied between 3 and 5 days per week. Water meters measured the amount applied to each irrigated replicate.
Water relations determinations.
Soil water content (SWC) was continuously monitored at depths of 20, 40, 70, and 110 cm with EnviroScan capacitance probes (Sentek Sensory Technologies, Stepney, Australia). Three access tubes per treatment in 2000 and 2001 and four in the remaining years were placed in the row line ~75 cm from a vine trunk and ~25 cm from a dripper. An in-situ calibration against volumetric soil moisture was previously performed by collecting undisturbed soil samples from each depth down to 100 cm in tubes installed for that exclusive purpose. A soil water balance was performed separately, on irrigated and nonirrigated vines, each year according to the capacitance probes readings. Total vine water consumption was estimated as the irrigation applied plus rainfall, plus or minus the variation in soil water content in the 10 to 120 cm soil profile. Runoff was considered negligible as the plots were nearly level and there was no visual evidence of it in most of the few storms that occurred during the study period. In year 2000, the capacitance probes were installed in late July, thus it was not possible to estimate the whole season vine soil water use for that year.
Determinations of water potential were performed with a pressure chamber (Soil Moisture Corp., Santa Barbara, CA) on four representative plants per treatment and two leaves per vine. The four vines were located in the same replicate, except in 2002 and 2003 when they were located in two replicates. Determinations were carried out at midday (1130 to 1230 hr solar) on bagged (stem water potential, Ψs) and uncovered leaves (leaf water potential, Ψl) at 2-week intervals. For the measurement of Ψl, leaf blades were covered with a moist cloth to minimize water loss during the measurement. Determinations of both Ψl and Ψs were not taken in the same vines as was the SWC.
Yield components.
Yield was determined at harvest on each of the seven internal rows (7 vines/row) of each replicate. The number of clusters per vine was determined in 12 vines per plot and average cluster weight determined from randomly selected samples of at least 20 clusters per plot. Berry weight was determined on random samples of about 200 berries per replicate.
Vegetative growth determinations.
Pruning weight (PW) and leaf area (LA) were determined in four vines per replicate. Leaf area was estimated after veraison when shoot growth had ceased. Leaf area per vine was estimated from a linear equation relating leaf area (Y, cm2 per shoot) and total (main plus laterals) shoot length (X, cm). This relationship was obtained from samples of about 10 to 20 representative shoots of different lengths collected after veraison each year. Thus, leaf area per vine was calculated from the sum of each of the measured individual shoot lengths. Leaf area to yield ratio (LA/Y) and yield to pruning weight ratio (Y/PW) were also calculated in the four selected vines per replicate.
Must and wine analysis.
Must components were determined in the same berry samples collected for berry fresh weight determination, which were crushed with a small handpress and the juice centrifuged. Soluble solids (Brix) were determined by refractometry. Juice pH and titratable acidity (TA) were determined by an automatic titrator. Organic acids (malic and tartaric in juice and wines and lactic only in wines) were analyzed by highperformance liquid chromatography following published procedures (Romero et al. 1993). Ethanol in the wines was analyzed by gas chromatography. Wine color intensity (OD420+OD520+OD620) and total phenolics index (OD280) were determined by spectrophotometry (Ribereau-Gayon et al. 2000) and were expressed in terms of absorbance units (AU). Anthocyanins (OD520 in HCl media) were also determined by spectrophotometry. All analytical determinations were duplicated.
Microvinification procedures.
Grapes from the different treatments were harvested on the same day (or with one day difference) when a minimum 21 Brix was reached and were transported to the experimental winery in field boxes. In 2000, nonirrigated vines could not reach the threshold Brix value and they had to be harvested at less than 21 Brix. Vinifications were performed at Estación Viticultura y Enología Requena separately on samples of ~30 kg from each plot, most often six vinifications per treatment. However, because of equipment and human resource limitations, vinifications could not be performed for all the experimental plots in all years. Grapes were mechanically crushed, destemmed, and fermented at ~25°C in stainless-steel containers. All wine lots were inoculated with a commercial yeast strain (L-2056; Danstar Ferment AC, Zug, Switzerland) at 100 mg/kg. Skin contact time was 7 days, and during this time they were punched down automatically every 4 hours. After alcoholic fermentation they were racked off and malolactic bacteria (Oenococcus oeni) inoculated. They were again racked off, sulfited at 100 mg/L K2S2O5, decanted, and bottled. Wine storage time was the same for all treatments and years. Analytical determination in the wines was performed at the same time in both treatments just before inoculation with malolactic bacteria, approximately one month after grapes were crushed.
Statistical analysis.
Analysis of variance was performed using the glm procedures of SAS statistical package (version 8.2; SAS Institute, Cary, NC). Data from each season were analyzed separately with irrigation as main factor. Across years, data were analyzed together with irrigation treatment, year, and their interaction as factors. Differences between treatment means were assessed by Dunnett’s t test against the nonirrigated vines (control). Simple linear regression analysis was carried out to explore relationships between parameters, and significance levels of the correlation coefficient at 5% or higher are reported. The slope of pairs of linear regressions were compared by means of the reg procedure of the SAS.
Results
Water relations.
The experimental period was characterized by two dry seasons, 2000 and 2005, while in the other years rainfall was about average for the area (Table 1⇓). As a consequence, at the beginning of 2000 and 2005, soil water content, even in the irrigated vines, was lower than the field capacity. Midday leaf water potential showed a decreasing tendency from the beginning of each season toward August; after which values tended to increase (Figure 1⇓). Midday stem water potential values followed a similar trend as leaf water potential determination but were in general 0.2 to 0.3 MPa higher (not shown). The minimum values of Ψl in rain-fed vines were reached during August and were ~-1.4 MPa except in the driest years (2000 and 2005) when the rain-fed vines reached Ψl -1.7 MPa. The lowest Ψl values of the irrigated vines during midseason were between -1.2 and -1.4 MPa.
Vegetative growth and yield.
Averaged over years, irrigation increased vine vigor, as indicated by the significant (p < 0.05) increase in pruning weight and the tendency to increased leaf area (Table 2⇓). Despite some scatter in the data, yield was linearly related to vine evapotranspiration (ET) (Figure 2⇓), with a slope of 0.044 t ha−1/mm ET. Averaged over 5 years, irrigation increased ET in 30%. However the effect of irrigation on ET varied greatly each year, with irrigation increasing ET by 12% in 2002 and by 71% in 2005 (Table 1⇑).
The average increase in yield due to irrigation was 31% (Table 3⇓), primarily because of increased berry weight, since cluster number and berries per cluster were only slightly and not significantly affected by irrigation (Table 3⇓).
In general, there were large differences among years in the leaf area to yield ratio, LA/Y (Table 4⇓). Irrigation only decreased this ratio in 2001 and 2005, and averaged over all seasons the effect of irrigation on LA/Y was not significant. Similarly, the Y/PW, also known as the Ravaz index, varied markedly among years (Table 4⇓), but it was not significantly affected by irrigation.
Must and wine quality.
Analytical values presented here are from must at harvest (Table 5⇓) and wines before malolactic fermentation (Table 6⇓). Overall, irrigation did not significantly affect must total soluble solids or wine alcohol content. However, the year-to-year effects of irrigation on those parameters were markedly different. In 2000 and 2004, when values of LA/Y where in general less than 1.0 m2 kg−1, irrigation increased the total soluble solids in must and, in 2000, the alcohol content of the wine. In 2005, when nonirrigated vines had much higher LA/Y values than irrigated vines, irrigation had the opposite effect, decreasing the Brix of the must and the wine alcohol content.
The effect of irrigation on the total acidity of the wine and the must was not significant (p > 0.05) (Table 5⇑, Table 6⇑). However, irrigation increased malic acid concentration but decreased tartaric acid in the wines. Wine and must pH were significantly (p > 0.05) higher in the irrigated vines.
Overall, irrigation did not significantly affect anthocyanin concentration and total phenolic content of the wines (Table 6⇑). However, irrigation reduced anthocyanin concentration and total phenolics in two (2001 and 2005) of the three years (2001, 2003, and 2005) of relatively low yield. A similar trend on the variable effect of irrigation among years could be observed for the color intensity of the wine. Nonetheless, irrigation decreased wine color on average over the 6 years (Table 6⇑).
Across years and treatments, wine alcohol decreased as crop level increased (Figure 3⇓). However, the effect of yield on wine alcohol content was more pronounced in the nonirrigated vines. There was a significant difference in the slope of the linear regression between yield and alcohol content (p < 0.001).
Discussion
The absolute values of SWC measurements were clearly not in agreement with leaf water potential measurements (Figure 1⇑). These results highlight the limitation of soil water content measurements in reflecting plant water status similar to other reports (Poni et al. 1994b, Intrigliolo and Castel 2006), but in contrast with studies in California (Williams and Araujo 2002), perhaps because SWC and plant water status were not measured in the same vines. This feature is particularly important considering the high spatial variability of SWC values that varied with year and ranked between a minimum value of the coefficient of variation of 21% observed in season 2000 and a maximum of 45% in 2002. It might be that the expected heterogeneity in the soil led to some errors in the calibration of the soil water sensor, a key concern in capacitance probes as their area of influence is small.
In our irrigated vines, plant water status tended to be lower than in previous reports in California with irrigated vines (Williams and Araujo 2002), reflecting that some degree of vine water stress occurred in our irrigated vines. In the first and last experimental years when rainfall was not enough to fulfill the soil water profile, some degree of water stress occurred in the irrigated vines even at the beginning of the season because irrigation did not start until June (Figure 1⇑).
Differences in water status between irrigated and nonirrigated vines were generally small except in the two dry years, 2000 and 2005. This was probably because the low amount of water applied to the irrigated treatment was insufficient to considerably influence vine water status.
In both irrigated and nonirrigated vines, yield varied widely between years showing a biennial bearing pattern. Across the years irrigation increased yield by 31%. This increase was slightly lower than that reported for Tempranillo grapevines in other regions of Spain (Esteban et al. 1999, García-Escudero et al. 1997). In Madrid, there was a reported increase of 40% in yield when irrigation was applied at 60% ETo (Esteban et al. 1999). In La Rioja, the yield increase attributed to irrigation was 40% when vines where watered at 48% of ETo. Irrigation also increased bud fertility in those studies, but in our study, this factor was not impaired by irrigation. The difference might be due to the soil characteristics of our site that allowed high water retention and thus a slow development of water stress (Figure 1⇑). Rain-fed vines reached considerably lower plant water status than irrigated vines late in the season (July–August), most likely after bud differentiation. Conversely, in other reports, even after veraison, water stress decreased bunch numbers in the next season (Matthews and Anderson 1989, Smart et al. 1974).
The effect of irrigation on yield was particularly noticeable in the final year, when vines experienced a similar water stress as the first year. However the increase in yield because of irrigation was much greater in 2005 than in 2000 (e.g., 149% and 19%; Table 3⇑), as irrigated vines in 2005 had more berries per cluster.
On average, irrigation had some negative effects on wine composition (Table 6⇑). It altered the balance between malic and tartaric acid, increasing the first and decreasing the latter. This consequently led to an increase in wine pH that, together with a slight decrease in anthocyanin concentration, reduced the color intensity. All these effects might be attributed to a dilution effect. Also, the higher vigor of the irrigated vines likely impaired cluster microclimate, reducing fruit light exposure.
Irrigation increased malic acid concentration by 33% and 25% in must and wine, respectively. Temperature is the main environmental factor affecting malic acid evolution and concentration in berries (Hale 1977). Irrigated vines had more vegetative growth, as reflected in greater pruning weights and leaf area (Table 2⇑). This larger canopy probably reduced cluster exposure to direct solar radiation and therefore cluster temperature—conditions favorable for the retention of malic acid and counteracting the dilution effect by irrigation because of larger berries. This observation agrees with findings by other authors, who have related acid content with temperature, and with the higher rate of malic acid degradation in nonirrigated vines because of less shading of the clusters by leaves (Smart et al. 1985).
Tar taric acid concentration decreased in irrigated wines, most likely because it is less affected than malic acid by environmental conditions (Ruffner 1982), and thus its concentration was probably more determined by the dilution effect and by increased precipitation of bitartrate potassium salts (Iland and Coombe 1988). Given that malic is a weaker acid than tartaric, the overall effect of irrigation on wine pH was to increase it. An increase in wine pH in response to irrigation has been previously reported (Freeman and Kliewer 1983).
Another negative impact of irrigation on wine quality was the decrease in wine color with an overall reduction of 18%, perhaps the result of increased wine pH and decreased anthocyanin concentration. An increase in pH might lead to lower fraction of pigments in the colored form (Freeman and Kliewer 1983). This negative tendency of irrigation on anthocyanin concentration and wine color is in agreement with other reports (Ginestar et al. 1998, Ojeda et al. 2002), but it is less pronounced than previous results in cv. Bobal in the same area (Salón et al. 2005). The only light effect of irrigation on wine quality was probably because irrigated and nonirrigated vines had similar levels of yield-pruning weight (Table 4⇑), suggesting that the balance between the vine demand and the supply was not drastically impaired by irrigation.
An increase in crop level led to a decrease in wine alcohol content. However, it is clear that the effect of increasing yield was different in irrigated than in nonirrigated vines (Figure 3⇑). Irrigation was able to mitigate part of the negative impact of higher yield on wine alcohol content. Thus, with yield values greater than 10 t ha−1, irrigated Tempranillo vines tended to have higher alcohol content in the wine.
The overall impact of irrigation on wine composition is also dependent on climatic conditions interacting with the yield values. Years 2000 and 2005 were both dry seasons, which led to a similar pattern of water status in the rain-fed vines (Figure 1⇑). However, yield was much lower in 2005 than in 2000. In 2000, irrigation helped to ripen the crop, as indicated by the lower soluble solids and alcohol content in the must and wine of the rain-fed vines. The opposite was found in 2005, with nonirrigated vines ranking higher in most wine characteristics, indicating that severe water stress, particularly if accompanied by high crop level values, can be detrimental to wine composition.
Conclusions
Irrigation increased vegetative growth and slightly improved vine water status, leading to an increase in vine evapotranspiration and consequently yield. This increase in yield was mostly compensated by the larger leaf area developed by the irrigated vines. Hence, irrigation did not impair the vine balance between sources and supplies expressed in terms of leaf area to yield and yield to pruning weight ratios. Wines made from irrigated vines had higher malic acid, lower tartaric acid, and increased pH, which might be detrimental to sanitary and aging stability of the wines made from the irrigated vines. Overall, wine phenolic content and anthocyanin concentration were not clearly affected by irrigation, and wine color decreased 18% in irrigated vines. Increasing yield linearly decreased wine alcohol content, but irrigation was able to mitigate, in part, the negative effect of increasing crop level. Supplying irrigation can be viewed as a tool to increase yield with a less detrimental impact on wine composition than under nonirrigated conditions.
Footnotes
Acknowledgments: This research was supported by funds from the Generalitat Valenciana, Consellería de Agricultura, Pesca y Alimentación, project 2002TAHVAL0034, and CiCYT, project 1FD1997-1276.
We are grateful to the Estación Experimental de Enología y Viticultura Requena for the vinifications and to the Servicio de Tecnología del Riego for meteorological data. Thanks are also given to C. Chirivella J.L. Salón, D. Pérez, S. Pedrón, E. Navarro, S. Cárcel, I. Yeves, C. García for help in field determinations and analytical data. The critical reading of the manuscript by A.N. Lakso and the English correction of M. Rose (Cornell University) are gratefully acknowledged.
- Received April 2007.
- Revision received August 2007.
- Copyright © 2008 by the American Society for Enology and Viticulture