Abstract
A study was conducted in a Chardonnay vineyard located in the Carneros district of Napa Valley to derive vineyard evapotranspiration (ETc) and seasonal crop coefficients (Kc) values. The vineyard was planted on 2.13 m rows, using a vertical shoot-positioned trellis. Vineyard ETc was measured using the soil water balance method. Soil water content (SWC) was measured in one-fourth of an individual vine’s soil profile (six access tubes per site) to a depth of 2.75 m. In addition, vines were irrigated with applied water amounts at 0.25, 0.5, 0.75, 1.0, and 1.25 of estimated vineyard ETc. Vineyard ETc the first year of the study was ~400 mm. Thereafter calculated vineyard ETc (the product of reference ET [ETo] and the Kc) ranged from 346 to 503 mm per season. Midday leaf water potential (Ψl), leaf net CO2 assimilation rate (A), and stomatal conductance (gs) were used to indirectly validate estimated ETc (to determine that vines were not stressed for water) and the derived Kc values. Midday Ψl, A, and gs were linearly related with applied water amounts and SWC across irrigation treatments and years. The diurnal measurements of A and gs resulted in differences among irrigation treatments, from early morning until late afternoon, with significant differences among treatments dependent upon actual applied water amounts. The results from this study are the first in which vineyard ETc has been measured on vines grown at a cool vineyard site in California. Estimates of ETc from this study would be valid for a vineyard with a row spacing of 2.13 m and a canopy vertically positioned using a maximum Kc of 0.74.
Vineyard water use (ETc) is dependent upon the age of the vine, the vine’s seasonal development, canopy size, row spacing, final use, and evaporative demand (Allen et al. 1998, Netzer et al. 2009, Williams 2010, 2012b, Williams and Ayars 2005a, 2005b, Williams et al. 2003a, 2003b). Other important aspects that determine vineyard water use are the amount and timing of precipitation events, the amount of water available in the soil profile, and irrigation practices (Williams et al. 2010a, 2012). Seasonal vineyard water use in areas of California where reference ET (ETo) exceeds 1400 mm during the growing season has been estimated (Williams 2012b) or measured (Grimes and Williams 1990, Williams and Ayars 2005a, Williams et al. 2003b) to range from 700 to 850 mm. Maximum seasonal ETc has also been shown to be higher than 1220 mm for grapevines trained to an open-gable trellis system for table grape production in Israel (Netzer et al. 2009). In Napa Valley, seasonal ETc was estimated at ~600 mm (Matthews and Anderson 1988). In Washington State, ETc values were much less (average of 417 mm per season) than ETc values in the San Joaquin Valley of California due to a shorter growing season and smaller canopies (Evans et al. 1993). Water use of grapevines grown in Spain was estimated to be 128 mm for dry-farmed vineyards (Aldaya et al. 2010), while in South Africa it ranged from a low of 226 mm (van Rooyen et al. 1980) up to 404 mm, the exact amount dependent upon the trellis system used (van Zyl and van Huyssteen 1980). The above would indicate that water use will vary considerably due to vineyard management practices (to include trellis type and row spacing), which are dependent upon final use of the grapes and evaporative demand and temperature during the season, which would be dependent upon vineyard location.
Grapevine water use (ETc) can be calculated as the product of ETo and the crop coefficient (Kc) (Allen et al. 1998). Crop coefficient values are for nonstressed crops cultivated under excellent agronomic and water management conditions and achieving maximum crop yield (i.e., standard conditions) (Allen et al. 1998). The primary objective of this study was to determine ETc of nonstressed, mature Chardonnay grapevines trained to a vertical shoot-positioned (VSP) trellis system when grown in a cool climate (Carneros district of Napa Valley) and to develop reliable, seasonal crop coefficients for such vineyards. During the first year of the study vines were irrigated based on evaporative demand and crop coefficients previously developed for grapevines (Williams et al. 2003a, 2003b) but adapted for the row spacing and trellis used in the experimental vineyard. Five irrigation treatments were imposed with applied water amounts at various fractions (0.25, 0.5, 0.75, 1.0, and 1.25) of estimated ETc for vines grafted onto two rootstocks (5C and 110R). Soil water content (SWC) was monitored in order to determine whether applied water amounts were adding to or subtracting from the soil water reservoir. If the soil moisture content remained constant, then the applied water amount was assumed to meet vineyard ETc (Grimes and Williams 1990, Williams et al. 2010a). A depletion of SWC would indicate that vine water requirements were not met and the irrigation amounts were less than ETc whereas an increase would indicate overirrigation.
The second objective of this study was to determine the effects of various applied water amounts, based on estimated ETc, on vine water status and leaf gas exchange. These physiological measurements have been used to indirectly validate the appropriateness of Kc values derived from ETc of non-water-stressed vines (Williams 2010, 2012b). For example, values of midday leaf water potential (Ψl) above a specific value (−1.0 MPa) are obtained from vines that are not stressed for water or fully irrigated (Williams and Baeza 2007). In addition, correlations among vine water status and soil water content were also examined to see how they compare with those made on a different cultivar grown in a hot environment (Williams 2012a). While this study was conducted for eight years, physiological measurements were taken at less frequent intervals during the final four years and only at the end of the growing season the final two years.
Materials and Methods
Location and plant materials.
The study was conducted from 1994 to 2001 in a Vitis vinifera L. cv. Chardonnay (clone 4) vineyard in the Carneros district (lat. 32°12′N; long. 122°21′W) of Napa Valley planted in 1990. The vines were grafted onto either 110 Richter (110R), a cross between V. berlandieri Ressequier No. 2 x V. rupestris Martin, or Teleki 5C (5C), a cross between V. berlandieri x V. riparia. The soil was a clay (51% clay, 36% silt, and 13% sand) and of uniform texture to a depth of 2.7 m. The soil bulk density was also uniform with depth and averaged 1.4 g cm3. The rows were oriented approximately east/west with a row spacing of 2.13 m and a vine spacing of 1.52 m (3086 vines/ha). All vines were trained to bilateral cordons and spur pruned. A vertical wire trellis system (VSP trellis) was used. Canopy management consisted of moving the wires and positioning the shoots. The shoots were hedged once they grew beyond the upper wires of the trellis (cut ~0.3 m above the uppermost wire). Leaves were removed within the fruiting zone on the north and south sides of the canopy up to the uppermost cluster on each shoot. The only clusters that were removed during the study were those developing on lateral shoots.
Grapevine phenology.
Vine phenology (budbreak, anthesis, and veraison) was monitored and estimated visually. Budbreak was considered to have occurred when green tissue was visible among the bud scales. The approximate dates of budbreak ranged from the third week in March to 1 Apr across seasons. The dates of anthesis ranged from 9 May in 1997 to 4 June in 1999 and the dates for veraison ranged from 2 July in 1997 to 5 Aug in 1999. The accumulation of degree days (>10°C) from date of estimated budbreak to harvest across the eight years of the study was 1244 (±15 se).
Treatments and experimental design.
The experimental design was a split-plot with randomized blocks. The main plots (rootstocks) were randomized within a block across rows, with each main plot consisting of 11 contiguous rows, or a total of 22 rows for the two rootstocks in each block. Blocks were replicated four times down all rows and were 17 vines in length. Two border vines separated blocks. Irrigation treatments were subplots within each main plot and consisted of water applications at various fractions (0.25, 0.5, 0.75, 1.0, and 1.25) of estimated vineyard ETc from the time irrigations commenced each season until the last irrigation of the year. The season-long application of water amounts at fractions less than ETc has been termed “sustained deficit irrigation,” or SDI (Fereres and Soriano 2007). Within each individual rootstock plot, the irrigation treatments were set up as a line source in which each row received more or less water depending on the direction the irrigation treatments within the plot were assigned. This design is similar to that used in previous irrigation studies by the author (Williams 2010, 2012b, Grimes and Williams 1990, Williams et al. 2010a). For example, if the first irrigation data row within a specific rootstock plot was irrigated at 0.25 of ETc, both rows bordering it were irrigated with the same amount of water. The next irrigation data row (proceeding to the right, the direction randomly assigned within each rootstock plot) would be irrigated with applied water at 0.5 of ETc. The border row to its left would be irrigated with applied water at 0.25 of ETc while that to its right would be irrigated with 0.5 of ETc. This procedure was then repeated for the three remaining irrigation treatments within each rootstock plot. Therefore, the row on either side of an irrigation data row was either irrigated with the same amount of water as the data row or with water amounts less than that designated for the specific irrigation treatment. Six border vines were used to separate an individual rootstock/irrigation treatment down the row: two vines from a row within a rootstock/irrigation treatment assigned to the preceding block and two vines in the next block and two border vines between blocks.
The irrigation treatments were changed subsequent to 1997. They were reduced to three treatments and consisted of vines receiving no applied water and those receiving applied water amounts at 0.5 and 1.0 of estimated ETc from 1998 to 2001 for both rootstocks. The no applied water treatment was established in the plots that were previously irrigated at 0.25 of estimated ETc.
Applied water amounts and soil water content.
Water was applied to the vines irrigated at estimated ETc (the 1.0 applied water amount) using two 4-L/hr emitters per vine, one on either side of the trunk. Vines irrigated at 1.25 of estimated ETc had two 4-L/hr emitters and one 2-L/hr emitter per vine (two emitters on one side of the trunk and one on the other). Those irrigated at 0.75 of ETc had one 4-L/hr emitter and one 2-L/hr emitter per vine (one on either side of the trunk). Vines irrigated at 0.25 and 0.5 of ETc had one and two 2-L/hr emitters per vine, respectively; the latter treatment had one emitter on either side of the trunk. Inline water meters in several rows measured actual applied water amounts for each irrigation treatment. The water meters were calibrated prior to the study.
Irrigation was initiated when midday leaf water potential (Ψl) of the 1.0 treatment was ~−1.0 MPa. Irrigation commenced 21 May 1994, 8 June 1995, 6 June 1996, and 19 May 1997 and was terminated on 15 Oct 1994, 20 Oct 1995, 9 Oct 1996, and 1 Oct 1997. Irrigation commenced on 30 June 1998, 10 June 1999, 15 May 2000, and 2 June 2001 and vines were irrigated until 9 Sept 1998, 4 Oct 1999, 19 Sept 2000, and 5 Sept 2001. During the 1995 growing season several problems occurred with the irrigation system. No water was applied for a four-week period from 28 June to 21 July due to pump failure and the time it took to repair. In the last week of July, too much water was applied for all treatments (equivalent to ~125 mm for the 1.0 irrigation treatment). Therefore, irrigation was terminated for a three-week period (from 2 until 24 Aug) that year.
Soil water content (SWC) for all irrigation treatments was monitored using the neutron back-scattering technique with a neutron moisture probe (model 503; DR Hydroprobe Moisture Gauge, Campbell Pacific Nuclear, Martinez, CA). Six access tubes were placed in one-fourth of an individual vine’s rooting volume and inserted to a depth of 3 m. Two access tubes were placed down the vine row (offset slightly from the drip line), one close to the trunk and one midway between vines within the row. Another two tubes were placed midway between rows, perpendicular to each of the two tubes placed within the row. The last two tubes were placed midway between the former two sets of tubes. Readings were taken at a depth of 0.23 and 0.45 m beneath the soil surface and then in increments of 0.3 m to a depth of 2.75 m. Each access tube site was replicated four times (one in each block) for vines of both rootstocks and irrigated at 1.0 of estimated ETc. Four access tube sites were installed in each of the other irrigation treatments, two per rootstock (in blocks 1 and 3). The neutron probe was calibrated according to Araujo et al. (1995) and SWC values expressed as percent by volume (θv). The SWC content at field capacity of this soil type was ~38% by volume and SWC at a soil moisture tension of −1.5 MPa was ~28% by volume. Therefore, total available water to a depth of 2.75 m for this soil at field capacity was estimated to be 275 mm. Soil water content was measured every two weeks during the growing season beginning the last week of March to first week in April. Measurements took place on Tuesday and the vines were irrigated Tuesday night to Wednesday by a research technician that week. Vines were irrigated the following week by an employee of the cooperator.
Determination of vineyard ET and seasonal crop coefficients.
A water budget was determined for each of the irrigation treatments from 1994 to 1999. The soil water balance can be calculated as:
where P is precipitation, I is irrigation amount, W is the contribution of a water table via upward capillary flow, ETc is vineyard evapotranspiration, R is surface runoff, D is drainage, and ΔSWC is the change in soil water content between measurement dates (Rana and Katerji 2000). Runoff was considered to be minimal and therefore neglected, as there were only two rainfall events throughout the six years where runoff would have been significant. The first occurred on 1 Apr 1996, with 55.6 mm rainfall, and the second occurred on 15 May 1996, with 49.5 mm rainfall. The first measurement of SWC in 1996 took place on 4 Apr, and therefore that rainfall event was not included in the water balance calculations. It was assumed that daily effective rainfall (in mm) could be calculated as (Prichard et al. 2004):
Therefore, effective rainfall on the second date (15 May, mentioned above) was calculated to be 34.5 mm. Soil water content was measured on 11 and 23 May 1996 and the increase in SWC was calculated to be 27 mm. Maximum daily rainfall amounts during the growing season across years exceeded 12 mm on only other two dates during the period in which SWC measurements took place. Therefore it was assumed that the frequency of SWC measurements (every two weeks) would provide an adequate measurement of rainfall contribution to the soil water reservoir.
The contribution of water from a water table was neglected. Open-ended access tubes were inserted to a depth of 3 m in the vineyard. Free water rarely was found in any of the access tubes, and therefore it was assumed that if a water table was present, then it was below (at least 3 m in depth) that where capillary upward flow would provide a significant portion of soil water availability (Stevens and Harvey 1996). Drainage was also assumed to be negligible in this study. If drainage had occurred, it was assumed that the depth to which SWC was measured would have identified the drainage; an increase in SWC with depth would have been detected. None was observed in this study, even for the 1.25 irrigation treatment.
Vineyard ET was calculated by rearranging Equation 1:
During the 1994 growing season, applied water amounts were added to the change of water in the soil profile of each treatment at approximately two-week intervals. The sum of both (ΔSWC was assumed to reflect any additions by rainfall) would be grapevine water use (ETc) for the designated interval (Grimes and Williams 1990, Williams et al. 2010b).
The crop coefficient (Kc) was calculated as:
where ETo is reference ET. Beginning in 1995, the Kc was calculated as a function of degree days (DD) from 1 Apr. Thereafter, ETc was calculated as:
The amount of shade cast on the ground at solar noon was measured on vines grafted to both rootstocks irrigated at the 1.0 level several times during the 1999 growing season. That value was converted to a Kc using an equation given elsewhere (Figure 10 in Williams and Ayars 2005b) and compared to the estimated Kc used in the study during that week. Reference ET was obtained from the California Irrigation Management Information System (CIMIS) weather station (#109) located ~6 km from the vineyard. Vines were irrigated once a week with the estimated required amounts for the 1.0 irrigation treatment and with fractional amounts of the 1.0 treatment for the other treatments.
Variables measured and calculations used to determine daily ETo can be found in Synder and Pruitt (1992). Temperature data used in calculating degree days were obtained from the CIMIS #109 weather station. Degree days were calculated using the sine method with a lower threshold of 10°C (for details, see the UC Statewide Integrated Pest Management Program website: www.ipm.ucdavis.edu).
Water potential and gas exchange measurements.
Grapevine water potential was measured according to the procedures of Williams and Araujo (2002). Predawn Ψ (ΨPD) measurements began at ≈0400 hr and were finished before sunrise using a pressure chamber (model 1000; PMS Instrument, Corvallis, OR). Midday measurements of leaf (Ψl) or stem (Ψstem) water potentials generally were taken between 1230 and 1330 hr Pacific daylight time. Leaf blades for ΨPD and Ψl determinations were covered with a plastic bag, quickly sealed, and petioles then cut within 1 to 2 sec. The time between leaf excision and chamber pressurization was generally <10 to 15 sec. Leaves chosen for ΨPD and Ψl were fully expanded and mature. Leaves chosen for Ψstem were taken from shoots on the shaded side of the canopy and measured with a technique described previously (Williams and Araujo 2002). For midday and diurnal Ψl measurements, leaves exposed to direct solar radiation at the time of measurement were used. A single leaf from a minimum of six to eight individual vine replicates (three to four from each rootstock) were measured and used for data analysis. It was determined early in the study that water potentials measured as a function of irrigation and/or rootstock treatments were similar in all four blocks of the study. This determination was also validated in subsequent years on several occasions. Therefore, most of the data presented in this paper were obtained in block 1. Midday Ψl was measured routinely the first four years of the study and infrequently in subsequent years.
Leaf net CO2 assimilation rates (A) and stomatal conductance (gs) measurements were measured as described by Williams et al. (2000) at the same time midday or diurnal Ψl measurements were taken. Both measures of gas exchange were made with a portable infrared gas analyzer (LCA2; Analytical Development, Hoddeson, United Kingdom) using the broad leaf chamber. Leaves chosen for gas exchange were similar to those used for Ψl in the same plots as mentioned above.
Statistical analyses.
Physiological data were analyzed via a two-way analysis of variance (ANOVA) with randomized blocks for data collected on the same date and/or time during the day using CoStat (ver. 6.400; CoHort Software, Monterey, CA). The first factor was rootstock and the second irrigation treatment. Means were separated using Duncan’s multiple range test and differences considered significant at p < 0.05. Data were also analyzed via regression analysis using linear, quadratic, and cubic terms. Regressions with the best fit are presented.
Results
While budbreak in this vineyard could begin the third week in March, degree day (DD) accumulation and estimated ETc started each year on 1 Apr. Degree days from 1 Apr to 31 Oct ranged from 1357 in 1999 to 1675 in 1997, with a mean of 1480 (Table 1). Reference ET (ETo) ranged from 885 to 1067 mm across years. Rainfall from November the previous year to October in the present growing season ranged from 253 to 904 mm for the 1994 and 1998 rainfall seasons, respectively.
Soil water content (SWC) measurements across years generally began the last week of March to first week in April and values usually ranged from 37 to 38 θv on those dates. In 1994 SWC values started to differ from one another subsequent to imposing the irrigation treatments on 21 May (day of year [DOY] 141) (Figure 1). Irrigation ceased on 15 Oct (DOY 288) after a rainfall event that year. The pattern of SWC across irrigation treatments throughout the 1996 and 1997 growing seasons was similar to that for 1994. Due to the failure of the irrigation pump in 1995 and the application of too much water later on, the pattern of SWC was not similar to that from the other years. Applied water amounts for the 0.25, 0.5, 0.75, and 1.25 irrigation treatments were 28, 51, 78, and 124% of those applied to the 1.0 treatment (Table 2). Applied water amounts in subsequent years for the various treatments were not always at the desired appropriate fraction compared to the 1.0 treatment. Water application amounts for the 1.25 treatment in 1994 maintained fairly constant SWC throughout the season compared to those irrigated with less water. While the SWC values of the 0.25, 0.75, and 1.25 irrigation treatments were similar on 9 June, water was depleted to a depth of 2 m on 29 Sept for all treatments, less so for the 1.25 treatment compared to the 0.25 and 0.75 treatments (Figure 2). Soil water content for the 0.25 irrigation treatment was also depleted to the 2.75 m depth that year. Soil moisture was greatest directly beneath the emitters and least midway between rows for all treatments at the end of the growing season (data not given). The SWC of the 0.25 treatment midway between rows was the lowest among all treatments.
During the 1994 growing season, comparisons of midday Ψl were made between the two rootstocks on each measurement date across irrigation treatments. There were several dates in which midday Ψl was significantly (p < 0.05) lower for vines grafted onto the 110R rootstock compared to the 5C rootstock by 0.03 to 0.05 MPa. The only significant interaction between rootstock and irrigation amount was on the last measurement date (26 Oct) in 1994. Midday Ψl values of vines (averaged across rootstocks) irrigated at 1.25 before the weekly irrigation event were never lower than −1.0 MPa in 1994, while the lowest Ψl for the 0.25 treatment reached −1.44 MPa on 21 Sept (Figure 3). Midday Ψl values of all treatments generally were higher than those measured the previous day after water was applied; however, values of the vines in their respective treatments remained distinct from one another. A similar pattern of seasonal midday Ψl was observed in 1996 and 1997 across treatments.
The SWC of the 1.25 irrigation treatment remained fairly constant throughout the growing season in 1994; thus it was decided to use the ETc data across rootstocks from that treatment for the calculation of the seasonal Kc using Equation 4 at approximately two-week intervals. The calculated Kc was expressed as a function of degree days from 1 Apr in 1994 and subsequent years (Figure 4). The Kc value began at ~0.1 and increased to 0.7 at the end of the growing season in 1994 and did not decrease at the end of the season. Estimated ETc for the 1994 growing season using Equation 4 expressed as a function of DD was only slightly higher (432 mm; Table 1) than that calculated using Equation 3 (401 mm) for the 1.25 irrigation treatment (Table 2). The shaded area under vines in the 1.0 treatment was measured several times during the 1999 growing season, converted to percent shaded area, and that value multiplied by 0.017 to provide an estimate of the Kc (see Materials and Methods). The equation used to fit the data would have a maximum Kc of 0.74 under the conditions of this study (2.13 m between rows and a VSP trellis).
Estimated ETc ranged from a low of 346 mm to a high of 503 mm in 1998 and 1997, respectively (Table 1). Estimated ETc as a percent of seasonal ETo ranged from 38% in 1999 to 47% in 1997. While differences of ETc among years were due to a combination of differences in ETo and DD from year to year, calculated ETc was more highly correlated with DD across years (ETc = −205 + 0.428*DD, R2 = 0.84, p < 0.0014, n = 8) than ETo. The greatest daily value of ETc averaged across years was 3.48 mm. Estimated ETc from 1 Apr to anthesis, 1 Apr to veraison, and 1 Apr to harvest averaged across all years was ~10, 38, and 78%, respectively, of season-long ETc. Differences in ETc among irrigation treatments ranged from a low of 227 mm for the 0.25 treatment in 1994 to a high of 554 mm for the 1.25 treatment in 1997 (Table 2). Seasonal ETc of the irrigated treatments from 1994 to 1999 was significantly correlated with applied water amounts (y = 201 + 0.653x, R2 = 0.83, p < 0.001). The ETc of the no applied water treatment was 260 mm in 1998 and 249 mm in 1999 (Table 2). The ETc of vines grafted onto 5C was 95% that of vines grafted onto 110R averaged across irrigation treatments and years.
Midday Ψl values measured during the growing season and averaged across years (1994 to 1997) were −1.18, −1.08, −1.01, −0.90, and −0.88 MPa for the 0.25, 0.5, 0.75, 1.0, and 1.25 irrigation treatments, respectively. The lowest Ψl values were reached at the end of the growing season, shortly before harvest (Table 3). The lowest midday Ψl values across years were recorded for vines in the no applied water treatments in 1998 and 1999. While the mean Ψl values were ~−1.85 MPa those two years, values as low as −2.0 MPa were measured at those times. Seasonal mean midday Ψl was significantly correlated with seasonal applied water amounts expressed in mm of water (y = −1.45 + 0.00267x + 0.00000327x2, R2 = 0.59, p < 0.001). Late season midday Ψl values were also significantly correlated with applied water amounts (data not given). Midday Ψl values were significantly correlated with SWC (Figure 5).
The relationship between leaf net CO2 assimilation rate (A) and stomatal conductance (gs) measured at midday was curvilinear (Figure 6). Leaf net CO2 assimilation rate did not appear to reach a maximum until a gs value of 450 to 500 mmol H2O m−2 s−2 was reached. Midday A and gs were significant (p < 0.001) linear functions of both SWC and midday Ψl, with R2 values ranging from 0.40 to 0.46 (data not given).
The lowest ΨPD values measured in this study were for the no applied water treatment in 1999. Values ranged from −0.5 to −0.6 MPa on several dates during August that year. Predawn Ψ values for vines irrigated at the 1.0 and 1.25 levels were generally no lower than −0.05 MPa. Diurnal measurements of leaf gas exchange and Ψl were made several times during the course of the study, with one representative date presented in Figure 7. Predawn Ψ values measured on that day were −0.18, −0.13, −0.07, −0.046, and −0.040 MPa for vines from the 0.25, 0.5, 0.75, 1.0, and 1.25 treatments, respectively. There were significant (p < 0.05) differences in ΨPD among each of the treatments on that day except for those of the 1.0 and 1.25 treatments. Minimum values of Ψl on that day occurred at 1500 hr across treatments, with the lowest value of −1.47 MPa recorded for the 0.25 treatment. Maximum values of A and gs occurred at 1300 hr and decreased slowly after that time.
Discussion
The degree days accumulated at the study site would indicate that this vineyard was located in a region I to II (cool to moderately cool), based upon the classification of Amerine and Winkler (1944) using a base temperature of 10°C. The degree days accumulated here averaged 1481 across years, which is only 56% the average degree days accumulated in the San Joaquin Valley of California (Williams et al. 2003b). While it might be assumed that evaporative demand would also be much lower at the Carneros location compared to the San Joaquin Valley, average ETo measured across years in Carneros was 86% of that reported in Williams et al. (2003b).
Vineyard ETc in this study was calculated using the soil water balance method. Others have used this method previously to calculate ETc in vineyards (Araujo et al. 1995, Lebon et al. 2003, Trambouze et al. 1998). The access tube arrangement used here is similar to elsewhere (Grimes and Williams 1990, Williams et al. 2010b), since it was assumed that the pattern of water in the soil profile under drip irrigation was not uniform (Mastrorilli et al. 1998), unlike that assumed for dry-farmed vineyards (Lebon et al. 2003). Therefore, SWC was measured in one-fourth of a vine’s soil profile using six access tubes per site to a depth of 2.75 m. This arrangement illustrated that the depletion of soil moisture at this location can occur midway between rows even if drip irrigation is used and to a depth of ~2.5 m, similar to that shown elsewhere (Williams and Trout 2005).
Calculated ETc the first year of this study was slightly higher than 400 mm, which is much lower than that measured on mature Thompson Seedless grapevines grown in the San Joaquin Valley (Williams and Ayars 2005a, Williams et al. 2003b). The lower values of ETc across years for this vineyard compared to vineyards in the San Joaquin Valley are due to a combination of effects. ETo was lower in Carneros than in the San Joaquin Valley, albeit by only 14%, but accumulated degree days were only 56% that measured in the San Joaquin Valley. Differences in estimated ETc among years in this study were predominately affected by the accumulation of degree days. It has been demonstrated that canopy development (as measured by the amount of shade cast on the ground at solar noon throughout the growing season) is highly correlated with degree days (Williams 2012b) and that the seasonal values of the Kc are a function of canopy development (Williams and Ayars 2005b). In this study, estimated ETc was 47% of ETo in 1997 but only 38% and 39% of ETo in 1998 and 1999, respectively. Degree day accumulations during both of the latter two years were only 81% of the 1997 growing season. Canopy type here was also different from the other studies (Williams and Ayars 2005a, Williams et al. 2003b), which used a California sprawl for Thompson Seedless. For the California sprawl, the amount of shade cast on the ground at solar noon could be up to 60%, whereas the maximum canopy that developed here using a VSP trellis shaded ~42% of the soil surface. Seasonal values of estimated and measured ETc in this study (350 to 500 mm) compare favorably with those of other studies in which a VSP trellis was also used (López-Urrea et al. 2012, Trambouze et al. 1998). Daily ETc values estimated or measured in this study at midseason (~3.5 mm) are comparable to those measured in other studies that used a VSP trellis (Lebon et al. 2003, López-Urrea et al. 2012).
The maximum Kc derived in this study (~0.74) was similar to that found for vines grown in a weighing lysimeter and trained to a VSP trellis (López-Urrea et al. 2012). In a different VSP trellised vineyard (Williams 2010), the maximum Kc was only 0.52 due to the fact that the row spacing was 3.05 m compared to the 2.13 m row spacing used here. The maximum Kc calculated for a Merlot vineyard located in the San Joaquin Valley with a California sprawl canopy was 0.7 (Williams 2012b). While the canopy for those vines would be much larger than the VSP trellised vines in this study, row spacing was 3.66 m and the percent shaded area of those vines only covered ~41% of the soil surface at midday compared to a maximum of ~43% of the soil surface at solar noon here. The seasonal Kc in this study did not decrease after midseason, as is often assumed (Allen et al. 1998). It has been demonstrated that grapevine Kc will not decrease as long as vines are irrigated with applied water amounts at ETc up to the end of October (Williams and Ayars 2005a) and the vines remain in a healthy state (Daane and Williams 2003). Once irrigation is terminated or irrigation amounts are reduced, the ETc / ETo ratio will decrease (Evans et al. 1993, Williams et al. 2012). The above would indicate a single seasonal Kc value as proposed for winegrapes (Allen et al. 1998) would not be appropriate in all vineyard situations where there are differences in row spacing, trellis type, and canopy size. However, I am of the opinion that the seasonal Kc developed for a particular row spacing and trellis (canopy) type would be appropriate across different locations if the Kc was calculated as a function of degree days from the date of budbreak or degree days from the same calendar date (close to the expected date of budbreak) each year.
Methods to assess vine water status should reflect the amount of water available in the soil profile (Williams and Trout 2005). Leaf water potential measured at midday in this study was highly correlated with applied water amounts across the first four growing seasons. Late season values of midday Ψl across all years of the study were significantly affected by irrigation treatment. In addition, midday Ψl was linearly correlated with SWC across years in this study. While the relationship between midday Ψl and SWC was slightly curvilinear elsewhere (Williams 2012a, Williams et al. 2012), a linear relationship would have fit the data equally well. Midday Ψl values measured here for the 1.25 and 1.0 irrigation treatments would indicate that applied water amounts were sufficient such that the vines were not stressed for water, similar to those reported elsewhere (Williams et al. 2010a).
The design of a study conducted in Paso Robles, California (Williams 2010) was similar to that described here with the same applied water amounts at specific fractions of estimated ETc. Midday Ψl values of Cabernet Sauvignon grapevines measured close to harvest across rootstocks and years in that study were −1.39, −1.23, −1.12, −0.96, and −0.93 MPa for the 0.25, 0.5, 0.75, 1.0, and 1.25 irrigation treatments, respectively. Midday Ψl values measured close to harvest during the first four years of this study were −1.4, −1.27, −1.16, −1.0, and −0.94 MPa for the same irrigation treatments, respectively. The similarity of Ψl values across irrigation treatments between the two studies would indicate that applied water amounts were based upon good estimates of ETc using reliable Kc values as a function of degree days and evaporative demand (ETo) specific for each vineyard location.
The relationship between midday measurements of A and gs in this study was curvilinear, with A maximized between gs values of 450 to 500 mmol H2O m−2 s−1. The highest values of gs measured in this study were slightly higher than 800 mm H2O m−2 s−1. While the shape of the curve relating A to gs in this study was similar to that reported for Thompson Seedless grown in a hot environment, maximum values of gs measured here were lower than those measured on Thompson Seedless (Williams 2012a), possibly due to cooler ambient temperatures of the Carneros district compared to the hotter temperatures in the San Joaquin Valley. In addition, gs of grapevine will significantly increase in response to increased ambient temperatures at a common vapor pressure deficit (Soar et al. 2009).
There were significant differences in Ψl among irrigation treatments on 5 Sept, whether measurements were taken before sunrise or later in the day. The lowest ΨPD measured on this date (−0.18 MPa for the 0.25 irrigation treatment) was similar to values others have measured on fully irrigated vines (Chaves et al. 2007, Correia et al. 1995, de Souza et al. 2005) or a value considered as no stress for field-grown grapevines (Schultz and Stoll 2010). Significant differences in A and gs among irrigation treatments later in the day would indicate that the vines on which the −0.18 MPa value was measured before sunrise were stressed for water. Predawn Ψ values similar to that measured for the 0.25 treatment here also reduced whole vine transpiration ~40% after irrigation was terminated (Williams et al. 2012). The aforementioned results would indicate that ΨPD may not be a reliable indicator of vine water status under drip irrigation when using RDI or SDI irrigation techniques, as values that are assumed to indicate little or no stress have been shown to significantly reduce vegetative and reproductive growth of grapevines (Williams et al. 2010a, 2010b).
The lowest values of Ψl measured on 5 Sept for the 1.25 and 1.0 irrigation treatments at 1300 and 1500 hr were higher than −1.0 MPa, a value considered to indicate little or no stress or values of fully irrigated vines (Williams 2012a, Williams and Trout 2005). The diurnal time-course pattern for A and gs in this study was similar to that reported for Thompson Seedless (Williams 2012a). Net CO2 assimilation rate and gs of Chardonnay irrigated at 0.25 of estimated ETc declined subsequent to 1100 hr, whereas those measurements for vines irrigated at 1.0 and 1.25 of estimated ETc increased up to 1300 hr before decreasing later in the day. Values of midday Ψl for vines irrigated at 1.0 and 1.25 of estimated ETc measured throughout the growing season, immediately before harvest, and on a diurnal basis would indicate that applied water amounts for those treatments were close to actual ETc.
Conclusions
Vineyard ETc in this study was determined using the soil water balance method. Estimated ETc ranged from 346 to 503 mm across years. The high and low values of estimated ETc were due predominantly to differences in degree days from one year to the next. Values of measured ETc in this study ranged from a low of 231 to a high of 560 mm of water across years and irrigation treatments. Rootstock had a minimal effect on ETc across years and irrigation treatment. The values of seasonal and daily ETc or the Kc were comparable to other studies in which a VSP trellis had been used or where percent canopy coverage was similar to this study.
Measures of vine water status, midday measurements of Ψl, A, and gs, were significantly correlated with applied water amounts and soil water content. The diurnal time course of Ψl, A, and gs indicated vines irrigated with various fractions of applied water amounts differ significantly from one another and that the depression in midday A and gs occurred only when the vines were irrigated with applied water amounts considerably less than full ETc. These results are similar to others obtained in a hot grapegrowing region in California. The above would indicate that vines grown in a cool grapegrowing region and irrigated with applied water amounts based upon good estimates of ETc, where the Kc is a function of degree days, respond similarly to vines grown elsewhere (warmer regions).
Acknowledgments
Acknowledgments: The research was funded in part by grants from the American Vineyard Foundation the first four years of the study. Mention of trade names or proprietary products is for the convenience of the reader only and does not constitute endorsement or preferential treatment by the University of California. The author thanks Peter Biscay, Weigang Yang, Nona Ebisuda, Jason Benz, and Paul Wiley for their technical assistance; the Robert Mondavi Winery, Daniel Bosch, Mitchell Kluge, and Don Williams for their assistance with this study; and Lawrence Schwankel and Mark A. Matthews for their helpful comments.
- Received August 2012.
- Revision received June 2013.
- Revision received October 2013.
- Accepted January 2014.
- Published online May 2014
- ©2014 by the American Society for Enology and Viticulture