Abstract
Relationships among vine water status, soil texture, and vine size were observed in four Ontario Pinot noir vineyards in 2008 and 2009. The vineyards were divided into water status zones using geographic information systems to map the seasonal mean leaf water potential (Ψ) and cane pruning weight (vine size). Leaf Ψ zones were confirmed using k-means clustering. Both seasons were cooler and wetter than average and the range of leaf Ψ defining the water status zones was narrow (−0.59 to −0.95 MPa across all vineyards). Yield, vine size, crop load, anthocyanins, and phenols had the highest coefficients of variability. Higher yields, berry weights, titratable acidity, anthocyanins, and color were occasionally associated with low water status zones. There were no berry composition variables with differences between vine size zones in all four vineyards. Higher yields, cluster numbers, and berry weights were frequently associated with high vine size zones. Principal components analysis separated the vineyards but did not create clusters based on leaf Ψ or vine size. There were notable correlations between vineyard and grape composition variables, and spatial trends were qualitatively related for many of the variables. Significant r2 values that suggested inverse relationships were found in 2008 for leaf Ψ versus anthocyanins, color intensity, and phenols and for vine size versus anthocyanins, while in 2009 there were significant r2 values for soil moisture versus anthocyanins and color intensity that likewise suggested inverse relationships. This study showed that there is potential for using geomatic techniques to understand variability in vineyards, but that erratic weather in eastern North America presents a challenge for understanding the driving forces of such variability.
- anthocyanins
- geographic information systems
- global positioning systems
- soil moisture
- terroir
- vine water status
In winegrowing regions, the effects that create differences between vineyards have been collectively referred to as terroir (Van Leeuwen and Seguin 2006). This idea can be applied to any product with characteristics that are unique to its region of origin, but is perhaps most associated with wine appellations of origin, which are renowned for their long history. There are many factors understood to be part of terroir that have been the subject of research around the winegrowing world. The regional climate, the site-specific mesoclimate, the soil pedology or texture, soil nutrient content and uptake by the vine, and the underlying geology of a region all play a role in defining terroir (van Leeuwen and Seguin 2006, Andres-de Prado et al. 2007, van Leeuwen 2010).
In younger regions such as the Niagara Peninsula in Ontario (ON), the degree of variation cannot be overestimated. In Niagara, there is a wide range of soil parent material, slope and aspect, distance from the moderating influence of Lake Ontario, and associated mesoclimate conditions (Shaw 2005). Soils are predominantly formed from parent materials based upon Halton clay till over Queenston shale and lacustrine sandy loam, with high water holding capacity (Kingston and Presant 1989). The Niagara Escarpment, the most prominent geological feature in the area, has exposed dolomite limestone cliffs with gentler slopes covered with silt and clay loams. These areas experience far better drainage and are almost entirely north-facing (Shaw 2005). This variability in soil characteristics can impact vine vigor, yield, and perhaps water status. A significant growth in the number of small artisanal wineries has permitted production of wines that are unique to individual vineyard sites, and in some cases, unique to specific vineyard blocks. In the past 10 to 15 years, this interest has expanded to include identification of unique portions of vineyard blocks, some <1 ha, that might be capable of producing extremely high-value wines based upon yield, vine size, or water status-based quality levels.
Physiological stress normally occurs when grapevine transpiration, governed by solar radiation, temperature, and relative humidity, exceeds available water (Hardie and Considine 1976). Water stress may reduce fruit set and yield (Hardie and Considine 1976), increase sugar accumulation and breakdown of malic acid (Koundouras et al. 1999), increase concentrations of anthocyanins and total grape phenolics (Sivilotti et al. 2005, Koundouras et al. 2006), and generally increase desirable grape composition and wine sensory attributes (Matthews et al. 1990, Reynolds et al. 2007, 2010, Willwerth et al. 2010).
The increasing use of irrigation in many New World vineyards makes it critical to understand how applying or withholding water from vines changes their growth habits and the composition of their fruit. Conversely, where irrigation is not used, the water status of the vines may be manipulated through other cultural practices, but will ultimately be affected by variations in the soil, with consequences for the composition of the fruit (Acevedo-Opazo et al. 2008, 2013). Variable water status within a vineyard is itself a component of the terroir effects of that site (van Leeuwen 2010). There is ongoing disagreement in the literature as to the effect of water stress on grapevine physiology and the characteristics and quality of the resulting wine. These disagreements may arise from other factors influencing vine growth, ultimately included in a broad definition of terroir.
In precision viticulture, there is a focus on understanding spatial and temporal variability in the production of wine-grapes (Hall et al. 2003, 2011). Grapegrowers have traditionally accepted variability within vineyards as inherent to the underlying qualities of the site itself, the terroir. Based on many years of experience, vineyard areas have been subdivided into individually rated vineyards of higher or lower quality. The emergence of geomatics software has allowed grapegrowers to geographically link information from their vineyards into the precision agriculture feedback loop and target inputs to specific regions of their vineyards. Analysis tools for potential use in precision viticulture have been assessed in New World regions including California (Johnson et al. 2001), Australia (Hall et al. 2003, Lamb et al. 2004, Bramley 2005, Bramley et al. 2001, 2011a, 2011b), and New Zealand (Trought and Bramley 2011, Bramley et al. 2011c), and in Old World regions such as Spain (Zarco-Tejada et al. 2005) and France (Acevedo-Opazo et al. 2008, 2013).
The main objective of this study was to assess geographic information system (GIS) tools to assist in understanding factors that contribute to the terroir effect of the Niagara winegrowing region of Ontario, Canada. Four Pinot noir vineyards in the Niagara Region were the study locations. This study encompassed several facets, including assessment of relationships between vine water status versus remote sensing variables (Ledderhof et al. 2016), as well as links between vine water status and wine sensory attributes (Ledderhof et al. 2014). It was hypothesized for this specific component of the study that temporally-consistent vine water status zones could be delineated that would be related spatially to yield components and berry composition. In particular, it was anticipated that berry pH, titratable acidity (TA), total soluble solids (TSS), anthocyanins, phenols, and color intensity would be favorably impacted in low vine water status zones of the study vineyards.
Materials and Methods
Vineyard blocks and sentinel vines
Four Vitis vinifera L. cv. Pinot noir vineyard blocks in the St. Davids area of Ontario [Coyote’s Run Estate Winery (three blocks) and Five Rows Craft Wine of Lowrey Vineyards] were chosen in 2008 (Ledderhof et al. 2014, 2016). Two blocks were in the Red Paw Vineyard (RP1, RP2) and one in the Black Paw Vineyard (BP) at Coyote’s Run in the Four Mile Creek sub-appellation. The fourth vineyard block was at Five Rows Craft Wine of Lowrey Vineyards (hereinafter, Lowrey) in the St. Davids Bench sub-appellation. Details on vineyard area, year planted, clone, rootstock, and row orientation are presented (Table 1). Vine × row spacing was 1.2 m × 2.4 m for all blocks. Vines were cane-pruned and trained using vertical shoot-positioning. All blocks had tile drainage in every other row.
A number of sentinel vines that were distributed evenly within each vineyard block were selected, flagged, and geolocated for repeated data collection. The panels at either end of the rows and the rows at the edge of each block were not used to select sentinel vines. A single sentinel vine was in every panel in every other row except at Lowrey, where sentinel vines were in every third row. The numbers of sentinel vines by vineyard block and per ha are listed (Table 1). Of these, subsets of one in five sentinel vines distributed throughout the vineyard blocks were selected for leaf water potential (Ψ) measurements. In total, there were 317 sentinel vines and 66 water status vines. Sampling strategy maps can be seen overlaid onto images of the vineyards (Figure 1). Sentinel vines were geolocated on 29 and 30 May 2008 using a Trimble GeoXT Handheld GPS running Trimble TerraSync 2.53 software (Trimble Navigation Ltd.) with ~8.6 m accuracy. Post-collection differential correction was performed using GPS Pathfinder Office 3.10 (Trimble Navigation Ltd.) to sub-meter accuracy using the Port Weller, ON base station correction. Final accuracy was 30 to 50 cm. The map projection used was in Universal Transverse Mercator (UTM) coordinates, Zone 17N with the 1927 North America Datum.
With the exception of harvest and pruning, all regular operations were carried out on the sentinel vines by the vineyard crews. This included pesticide applications, mid-season hedging, soil tilling, and cluster-thinning. All cultural practices were consistent with those recommended (Ontario Ministry of Agriculture and Food 2008).
Spatial mapping
All field and berry sample measurements were tied to specific vines, so GIS software was used to map the variables onto a two-dimensional surface. Parameters were mapped using Surfer (Version 8.05, Golden Software, Inc.). Data were gridded using the Modified Shepard’s Method (Shepard 1968, Renka 1988a, 1988b). This method is similar to the inverse distance weighting (IDW) method, but uses a local least squares method to eliminate the bull’s-eye effect created by extreme values. The Modified Shepard’s Method was made a smoothing interpolator with the inclusion of a 0.2 smoothing factor. Variation was assumed to be isotropic and a round search radius was used for gridding.
The grid line geometry was determined independently for each block. X and Y direction max and minimum values were extended by several meters to create a rectangular frame around the vineyard block without any sentinel vine touching the edge of the grid. The larger direction was assigned 100 lines by default, and in the other direction, the number of lines was assigned to keep the grid blocks as close to square as possible. The sizes of the grids, X × Y in meters, were RP1: 1.82 × 1.96; RP2: 1.31 × 1.43; BP: 2.20 × 2.01; and Lowrey: 1.74 × 1.85. Since grid node values were determined by the surrounding nodes, those at the extents of the maps were often assigned unreasonable values. A blanking file was created for each vineyard to isolate the sentinel vines within the larger vineyard map and eliminate the extreme values. Where unreasonable values occurred inside the vineyard block (such as a negative value for yield), grid math was used to replace these values (for yield, negative values were replaced with zeros). The extents of the color scale were adjusted for each map, which must be considered when comparing maps of the same variable between vineyards or across vintages.
Soil sampling
Soil samples were collected at each water status vine on 22 to 26 May 2008. Samples were taken from north of the vine trunk (west for RP1). A single gauge auger (Eijkelkamp Agrisearch Equipment) was driven vertically to a final depth of 75 cm, then the entire core was homogenized and shipped to Agri-Food Labs (Guelph, ON) for analysis of soil pH; buffer pH (when pH <6.8); organic matter (OM; %); P, K, Mg, and Ca (mg/kg); cation exchange capacity (CEC, mEq/100 g); and texture (% silt, sand, and clay) using standard procedures (Canadian Society of Soil Science 1993).
Soil moisture
Soil moisture was measured by time domain reflectometry (TDR) using the Field Scout model TDR 300 Soil Moisture Meter (Spectrum Technologies) fitted with a pair of 20-cm stainless steel probes. The volumetric water content mode was used for the 20-cm depth, using the high clay setting for soils with >40% clay content. Measurements were made bi-weekly for all sentinel vines in both 2008 and 2009. In 2008, seven sets of measurements were collected on 19 June; 2, 14, and 31 July; 12 and 27 Aug; and 8 Sept. In 2009, six sets of measurements were collected on 8 and 20 July; 5 and 19 Aug; and 3 and 17 Sept. In each case, where possible, there was at least 24 hr between the last rainfall event and data collection. The first two measurements were taken on alternate sides of the trunk, within 30 cm of the vine. If the two measurements differed by more than 10%, a third measurement was taken roughly at the midpoint between the first two measurements. The two or three measurements were averaged for a single value for each vine for that date.
Vine water status
Midday leaf Ψ was measured using a pressure chamber (Soil Moisture Equipment Corp.) and the technique described (Turner 1988). The measurements were made only on the subsets of sentinel vines selected for this purpose, between 1000 and 1400 hr on the same days as soil moisture was measured. For each vine, two fully expanded, fully exposed leaves from different parts of the vine were measured. If the readings differed by >0.15 MPa, then a third leaf was measured. Each leaf was excised with a razor blade transverse to the length of the petiole and immediately inserted through the lid of the pressure chamber with the cut end exposed. N2 gas was used to slowly pressurize the chamber until sap began to flow out of the cut end of the petiole and the pressure was thereafter recorded. Vine water status zones were delineated based on seasonal means of all pressure bomb measurements. Using the previously described mapping techniques, Ψ maps were created for each block and divided into zones. The threshold value dividing high- and low-leaf Ψ zones was based on the median value, such that the number of vines in each zone was roughly equal. Since the range of leaf Ψ values was different in each vineyard block, a different threshold value was used to divide each block. In 2008, the RP1, RP2, and BP vineyards were divided into high and low water status and the Lowrey vineyard was divided into high, medium, and low water status. In 2009, all four vineyard blocks were divided into high and low water status. These divisions can be seen for the RP1 (Figure 2), RP2, BP, and Lowrey vineyards (Figure 3), respectively. k-Means clustering for these water status zones is also shown (Figure 4).
Vine size
Vine size was measured as dormant cane pruning weight. Timing of pruning and the number of nodes retained per cane was determined by the winery/grower. For the 2008 season, the Lowrey vineyard was pruned on 14 Dec 2008 and the RP1, RP2, and BP vineyards were pruned on 17 Feb 2009. For the 2009 season, the Lowrey vineyard was pruned on 15 Dec 2009. The RP1, RP2, and BP vineyards were pruned by the winery’s field crew in early February 2010, before the sentinel vines could be pruned. In both years, the Lowrey vineyard was pruned to two canes with 10 to 12 nodes each. In the 2008 season, the RP1, RP2, and BP blocks were pruned to three canes with 10 to 12 nodes each. The dormant prunings were weighed in-situ using an electronic field scale (RSDS-50; Rapala).
Harvest and yield data
Harvest dates were at the discretion of the vineyard managers. In 2008, the Lowrey vineyard was harvested on 16 Sept, RP1 and RP2 on 29 Sept, and BP on 30 Sept. In 2009, RP1 was harvested on 1 Oct, Lowrey on 5 Oct, and RP2 and BP on 6 Oct. All fruit from the sentinel vines was harvested, weighed, and the number of clusters per vine was counted. Mean cluster weight was calculated from these data. Fruit to be kept for winemaking was bulked by water status zone.
Berry composition
Sample preparation
At harvest, a randomly selected sample of 100 berries was taken from each sentinel vine and frozen at −25°C until further analysis. The berry sample was weighed to determine the mean berry weight and then placed in a 250 mL beaker in a water bath at 80°C for 1 hr to dissolve all precipitated tartaric acid. The samples were allowed to cool and then homogenized in a commercial juicer (Model 500; Omega Products). After settling, juice was decanted from the top layer of foam.
Soluble solids, pH, and titratable acidity
TSS was measured as Brix using an Abbe benchtop refractometer (Model 10450; American Optical). Berry pH was measured using an Accumet pH/ion meter and VWR SympHony electrode. Juice samples (~35 mL) were clarified by centrifugation at 4500 g for 10 min using a Centra CL2 benchtop centrifuge (International Equipment Co.). The remainder of the juice (~20 mL) was placed in plastic snap-top vials and returned to the −25°C freezer for subsequent color analysis. TA was measured on 5 mL of centrifuged juice, titrated to an endpoint of pH 8.2 with 0.1 N NaOH using a PC-Titrate autotitrator (PC-1300-475; Man-Tech Associates).
Color/hue, total anthocyanins, and total phenols
Samples were heated at 80°C for 30 min, then centrifuged at 3500 g at 4°C in a refrigerated centrifuge (Model B-20; International Equipment Co.) before analysis of color/hue, total phenols, and total anthocyanins. Color and hue were measured using a modification of a previously reported method (Mazza et al. 1999). In 2008, samples were loaded directly into a 1-mm path length quartz cuvette. Samples were darker in 2009, so they were diluted 1:10 in 9 mL of pH 3.5 buffer (0.1 M citric acid + 0.2 M Na2HPO4) and thereafter read in a 10-mm path length plastic cuvette. In both years, absorbance at 420 nm and 520 nm was measured using a UV-vis spectrophotometer (Ultrospec 2100 Pro; GE Healthcare Life Sciences). Color intensity was calculated as A420+A520 and hue, as A420/A520. Total anthocyanins were quantified using the pH shift method (Fuleki and Francis 1968). Samples were diluted 1:10 in 9 mL of pH 1.0 buffer (0.2 M KCl and 0.2 M HCl) and pH 4.5 buffer (1 M NaOH and 1 M HCl) and mixed by vortexing. The samples were allowed to sit in the dark for 1 hr to equilibrate. In a 10-mm path length plastic cuvette, absorbance at 520 nm was measured using a UV-vis spectrophotometer. A standard curve was generated using six concentrations of malvidin-3-glucoside. Total anthocyanins were calculated as (A520, pH1.0-A520, pH4.5)/0.0042, in mg/L malvidin equivalents.
Total phenols were quantified using the Folin-Ciocalteu micro method (Waterhouse 2001) based on Slinkard and Singleton (1977). A calibration curve was created, with each set of samples evaluated using a 5000 mg/L stock solution of gallic acid (0.5 g gallic acid in 10 mL ethanol, brought to a volume of 100 mL with water). Gallic acid concentrations in the standard curve were 0, 50, 100, 150, 250, and 500 mg/L. Samples were diluted 1:10 in 9 mL of distilled water in test tubes and mixed by vortexing. Twenty μL of each sample or standard was pipetted into a 10-mm path-length plastic cuvette, to which 1.58 mL of water was added. Thereafter, 100 μL of the Folin-Ciocalteu reagent (VWR Scientific) was added to each cuvette, followed by mixing. After 30 sec but no longer than 8 min later, 300 μL of 20% anhydrous NaCO3 was added to the cuvettes with mixing. Solutions were left in the dark for 2 hr at room temperature. Absorbance at 765 nm was measured using a UV-vis spectrophotometer. Total phenols were determined from the standard curve, corrected for the dilution in water, and expressed in mg/L gallic acid equivalents.
Data analysis
Gross variation of yield components, grape composition, and vineyard variables was analyzed using previously described methods (Bramley 2005). The median and coefficient of variation (CV) were calculated to express distribution of the data points. Within each vintage, the range (max and min values) was used to express the variation of each variable within a vineyard. Spread was calculated from the range divided by the median, expressed as a percent, which is an indicator of degree of variation in each variable that is potentially of greatest value in an industrial context; it is a normalized value that can be used to compare variation across variables and vintages (Bramley 2005). Analysis of variance (ANOVA) was performed on data from each vineyard block by vintage. Sentinel vines were grouped first by water status zone and alternatively, by vine size zone. Separate ANOVAs were conducted to determine whether vine and fruit characteristics differed among water status categories and between vine size categories using the GLM procedure in SAS (Version 9.1.3; SAS Institute, Inc.), with means separation by least significant difference. Pearson’s correlation matrices were generated using the CORR procedure for all variables measured on sentinel vines by vineyard. Principal component analysis (PCA) was performed on the mean values grouped by water status zone for all vineyards using JMP (Version 8.0.1; SAS Institute, Inc.). k-Means clustering was used to determine the veracity of determination of the water status zones and to verify the accuracy of the interpolation.
Results
Within-block differences
The within-vineyard gross variability of yield components, berry composition, and vineyard soil variables, including soil moisture and leaf Ψ, are reported in Supplemental Tables 1 to 4 for RP1, RP2, BP, and Lowrey vineyards, respectively. In all four vineyards, in both years, berry pH had the smallest CV and spread, followed by hue and TSS. These three berry composition measurements had the least gross variability within each vineyard block. Crop load, vine size, and yield had the highest degree of gross variation within each vineyard. Anthocyanins and phenols also had high CVs and spread. It is notable that for these two berry composition metrics, there was more variability in 2008 than in 2009 in all four vineyards, more so than was observed in any other metric. In terms of soil variables, while soil texture at each block was predominantly clay, the sand component was the most variable. With the exception of RP1, where % clay had the lowest CV and spread, the other three blocks were least variable in soil pH.
Yield components and berry composition
Vine water status
Grouped by water status and vine size categories, means for soil, vine, and fruit characteristics of sentinel vines are presented (Tables 2 to 6). Division by water status zone was verified by the highly significant difference between leaf Ψ category means in each block. There were no variables for which there were consistent differences between water status zones at all four blocks, in either vintage. In 2008, cluster weight, berry TA, and color intensity were different between water status zones in three of the four vineyards; however, for each of these metrics, the direction of the trend was not the same for all three vineyards. The low water status zone had higher TA in the BP vineyard, but the high water status zone had the higher TA in RP2 and Lowrey vineyards. In 2009, there were never more than two of four vineyards with differences between water status zones.
Considering both vintages, vine size was only associated with water status in one instance, in which smaller vines were associated with lower water status in the Lowrey vineyard in 2009 (Table 2). Soil moisture was slightly lower in low water status zones in RP1 (2009), but inexplicably highest in the low and medium zones in Lowrey in 2008. Among yield components, RP1 had higher yields (2008 and 2009), cluster number (2009), and cluster weight (2008) in low water status zones, while RP2 had higher berry weights (2009), and Lowrey had highest cluster weights (2008 and 2009) in low water status zones (Table 3). TSS was unaffected by water status, but TA was higher in low water status zones in three instances (RP1 2009, BP 2008, and 2009) and lower in two others (RP2 2008, Lowrey 2008). Berry pH was lower in low water status zones in two instances (RP2 2009, BP 2008) (Table 4). Anthocyanins and color both increased in low water status zones at RP2 and Lowrey in 2008; however, anthocyanins decreased in the low water status zones at RP2 in 2009 and color decreased at RP1 in 2008. Phenols were higher in low water status zones in RP2 in 2009 (Table 5). Soil texture variables were infrequently related to vine water status: low water status zones at the Lowrey site had higher % clay (both vintages) and lower % sand (2009), and also had higher % OM (2009) and CEC (both vintages). Low water status zones in RP1, on the other hand, had lower % OM and CEC and lower pH in 2008, while BP displayed higher % OM and CEC in 2009 (Table 6).
Vine size
Means by vine size category for soil moisture, leaf Ψ, and vine size for all four blocks in both years are shown (Table 2), as are means for yield components (Table 3), berry composition means (Tables 4 and 5), and soil analysis variables (Table 6). Division by vine size status zone was verified by the highly significant differences among category means for pruning weight in all blocks in 2008 and at Lowrey in 2009. There were no variables that differed between vine size zones in all four vineyards in 2008. In 2008, berry weight and % sand were different between vine size status zones in three of the four vineyards. The berry weight trend was the same in each of those three vineyards, with the high vine size status zone having larger berries. The trend in % sand was not consistent over all three vineyards. In 2009, only the Lowrey block was evaluated by vine size. Vine size zones were different for all yield components, berry TA, color intensity, and % clay, sand, and % OM.
Considering both vintages (four blocks in 2008 and one in 2009), lower soil moisture was unexpectedly associated with high vine size in two instances (Table 2), although this may be attributable to higher water demand in larger vines. Yield and clusters per vine were highest in high vine size zones in three of five instances, while berry weights were highest in four of five instances (Table 3). TSS and pH increased in high vine size zones in three instances (RP2 in both years and Lowrey in 2008), and TA was highest in high vine size zones in 2009 (Table 4). Anthocyanins were lowest in two instances in high vine size zones (RP1, BP), as was color (RP1), but color was higher in high vine size zones in Lowrey in 2009 (Table 5). Zones with higher vine size had lower % clay in three of five instances, lower % silt (RP1), and higher % sand in three of five instances (Table 6). In 2009, Lowrey also had lower % OM and CEC in high vine size zones.
Spatial analysis
Maps were used to visualize spatial variability in vineyard and vine characteristics. Leaf Ψ and water status zone delineation for both vintages are shown for RP1 (Figure 2) and the other three vineyards (Figure 3). Spatial variation in water status zones, the basis for vineyard intra-block divisions in this study, was somewhat stable between vintages. RP1 had the lowest values in the western half of the block in both years, with a branch extending east and north in 2009 cutting through the same region as the low water status zone in 2008 (Figure 2). In RP2, there was a zone of lower water status through the middle of the vineyard in both years and the two maps were very similar (Figure 3). BP water status zones were roughly similar between years, with the high water status zone running through the middle of vineyard from north to south (Figure 3). Three water status zones were delineated in the Lowrey vineyard in 2008, while there were two in 2009 (Figure 3). Ignoring dividing lines between zones, the lower water status zones were located in the north end of the vineyard in each year.
The threshold value dividing high and low leaf Ψ zones was based on the median value for each vineyard. The veracity of this approach was tested using k-means clustering (Figure 4). At RP1 in 2008, four low leaf Ψ vines were assigned to the high leaf Ψ zone, but the low leaf Ψ zone contained exclusively low leaf Ψ vines. In 2009, all high leaf Ψ vines were assigned to the high leaf Ψ zone and all low leaf Ψ vines were likewise assigned to the low leaf Ψ zone. For RP2 in 2008, two low leaf Ψ vines were assigned to the high leaf Ψ zones and three high leaf Ψ vines were assigned to the low leaf Ψ. In 2009, one low leaf Ψ vine was assigned to the high leaf Ψ zone, but the low leaf Ψ zone contained only designated low leaf Ψ vines. For BP in 2008, one low leaf Ψ vine was assigned to the high leaf Ψ zone, while the low leaf Ψ zone contained low leaf Ψ vines only. In 2009, the two zones contained only vines designated to those zones. In 2008, the high leaf Ψ zone at Lowrey contained all high leaf Ψ vines, while the low leaf Ψ zone contained all but one of the low leaf Ψ vines plus two medium leaf Ψ vines. In 2009, the high leaf Ψ zone contained all high leaf Ψ vines, while the low leaf Ψ zone contained all low leaf Ψ vines plus three high leaf Ψ vines. These results suggest that the interpolation process used for map creation was for the most part accurate, but that there was error nonetheless at the level of individual vine assignment to vine water status categories.
A visual comparison of maps of all other variables measured in the four vineyards (Supplemental Figures 1 to 12) allowed assessment of the spatial variability and relationships among vineyard, vine, and fruit characteristics. In both years, similarities between maps of yield and cluster weight showed an association between these variables. Across vintages, some trends were present in both years and others were not consistent. At RP1 in 2008, there was a distinct band of higher yield, running north-east through the eastern half of the vineyard (Supplemental Figure 5). This region was still present in 2009, although yields were slightly lower (Supplemental Figure 6). Within the same vintage, relationships between berry composition variables were similar and there were clear patterns of higher TSS corresponding with higher pH and lower TA. Comparing the same variables in 2009, there were some similarities between years, but also some differences. There was a region of higher pH in the eastern half of the vineyard in both years. The western half of the vineyard had lower TSS in 2008, but this pattern was reversed in 2009. BP maps were more difficult to interpret due to block geometry. This block was very narrow compared to its length, which made surface interpolation difficult (Supplemental Figures 9 and 10). Spatial patterns in anthocyanins, color, hue, and phenols were very similar to one another within vineyards and within vintage (Supplemental Figures 5 to 12). Supplemental Figures 11 (2008) and 12 (2009) show this for the Lowrey vineyard. Between years, these spatial trends also appeared relatively stable: in the western half of the Lowrey vineyard, the low anthocyanins zone in 2008 was present again in 2009.
Linear correlation and regression
Linear correlation coefficients between yield components, berry composition, vine size, and soil metrics for the pooled data from all vineyards are shown for 2008 (Table 7) and 2009 (Table 8). There were very significant correlations between yield and both cluster number and cluster weight in both years. Yield also correlated inversely with berry anthocyanins and color in 2008, and berry weight correlated strongly with berry pH and TA and inversely correlated with mean soil moisture. Leaf Ψ correlated inversely with berry pH, TSS, vine size, berry anthocyanins, color, and % clay, and directly correlated with % sand in 2008. Soil moisture correlated with color, phenols, % clay, % silt, soil CEC, and soil pH, while vine size correlated with leaf Ψ and inversely correlated with berry pH and anthocyanins. Berry weight correlated with hue and % silt and inversely correlated with color and % clay in 2009. Leaf Ψ correlated with hue and inversely correlated with % clay and % sand. Soil moisture correlated with % clay, soil CEC, and pH, and inversely correlated with anthocyanins and % silt. Vine size was measured only at the Lowrey site in 2009 and was not included in the analysis.
Regression analyses detected weak inverse relationships in 2008 between leaf Ψ and berry anthocyanins, color intensity, and phenols (Figure 5A to C), and between vine size and berry anthocyanins (Figure 5D). No relationships were detected between vine size and either color or phenols (Figure 5E and F). In 2009, no relationships were detected between leaf Ψ and berry anthocyanins, color, or phenols (Figure 6A to C), but soil moisture was inversely related to berry anthocyanins, color, and phenols (Figure 6D to F).
Principal component analysis
Principal component analysis of yield components, grape composition, and vineyard variables when grouped by water status zone and observation loadings are shown for 2008 (Figure 7) and 2009 (Figure 8). All four vineyards from both vintages are also shown (Figure 9). Two PCs explained 86.5% of the variation in vineyard, vine, and fruit characteristics at the four vineyard blocks in 2008 (Figure 7). With the exception of TA, all variables were heavily loaded on these components. Vine size and mean leaf Ψ (absolute value; a.v.) were highly correlated with % sand and inversely correlated with TSS, TA, anthocyanins, and berry pH. Yield correlated inversely with % clay, phenols, and color. The four vineyard blocks clustered in the observations plot, with some differences between water status zones within vineyards, but far larger differences between vineyards. BP was described by % clay, CEC, soil pH, OM, TSS, anthocyanins, phenols, and color; Lowrey by % sand, leaf Ψ, vine size, and cluster weight; RP1 and RP2 by crop load, cluster number, and berry weight. Vineyards clustered by block and there was no apparent separation based upon either leaf Ψ or vine size.
The first two PCs explained 70.3% of the variation in the 2009 data (Figure 8). There were more variables that were not heavily loaded on PC1 or PC2 than in 2008. Yield and cluster weight correlated inversely with anthocyanins, hue, and TSS. CEC, % clay, and soil moisture correlated inversely with mean leaf Ψ (a.v.), berry weight, and % silt. In 2009, all four vineyards separated in the observation loadings plot, but once again, there was no apparent separation based upon leaf Ψ. BP was loaded with % clay, CEC, OM, soil pH, and soil moisture. RP2 was loaded with leaf Ψ, TSS, cluster weight, berry weight, anthocyanins, and hue. RP1 was loaded with % sand, color, phenols, and TA. Lowrey was most closely loaded with yield, cluster number, and color intensity.
The PCA of data from both vintages explained 61.5% of the variation in the data with the first two components (Figure 9). Anthocyanins, phenols, TSS, and berry pH were all highly correlated. These metrics all roughly inversely correlated with yield, cluster weight, leaf Ψ (a.v.), and % sand. Neither % clay nor soil moisture correlated well with yield or berry composition metrics except for berry weight, TA, and color. One large group containing RP1 and RP2 (2008 and 2009) and Lowrey (2009) was located in the upper left quadrant and associated with % silt, berry weight, and TA. Lowrey 2008 samples were in the lower left quadrant and characterized by % sand, yield, cluster number, cluster weight, and leaf Ψ, while two groupings consisting of the BP samples were located along PC1 to the right of the others and were associated with several soil variables (% clay, % OM, CEC, and pH), soil moisture, color, and to a lesser extent TSS, pH, anthocyanins, and phenols. As with the 2008 and 2009 PCA, there was no apparent separation based upon leaf Ψ.
Discussion
Vineyard variability
The fundamental goal of this project’s hypothesis was to demonstrate significant and temporally-consistent spatial variability in leaf Ψ, vine size, and soil moisture, and to ascertain the existence of relationships between leaf Ψ and various yield components and berry composition characteristics. Spatial maps of leaf Ψ indicated the presence of temporally-consistent water status zones in the four vineyards. The magnitude of variability in water status and other variables can be normalized using “spread” for ease of comparison (Bramley 2005), where those variables with the highest CV are those with the highest spread. Thus, spread is a potential tool to convey how successfully a vineyard block has achieved consistency (Bramley 2005). In this case, the degree of variability in leaf Ψ was vineyard-specific, with CV values <10% under most circumstances and spread figures of 14.7 to 18.5% (BP), 15.4 to 26.3% (RP2), 32 to 47% (Lowrey), and 75.1 to 83.7% (RP1) (Supplemental Tables 1 to 4). Variability in vine size was substantially greater and CV and spread ranged from 39.0 to 57.2 and 200.8 to 261.5%, respectively (Supplemental Tables 1 to 4). Variability in soil moisture was similar across all vineyards and CV and spread ranged from 10.1 to 17.6% and 44.1 to 74.8%, respectively.
Water status zones in vineyards have previously been delineated in Ontario vineyards using leaf Ψ measurements and GIS (Reynolds et al. 2010, Willwerth et al. 2010, Reynolds and Hakimi Rezaei 2014a, 2014b, 2014c), and in some instances, these zones were related spatially to Normalized Difference Vegetation Index values obtained by remote sensing (Reynolds et al. 2010, Ledderhof et al. 2016). Stem Ψ was used to delineate water status zones in Bordeaux vineyards (van Leeuwen et al. 2006, 2009), and water status zones were demarcated using predawn leaf Ψ (Taylor et al. 2010). Definition of vigor zones have been widely described, including those in Pinot noir (Cortell and Kennedy 2006, Cortell et al. 2007a, 2007b, 2008), in Syrah in southern France (Tisseyre et al. 2008), in Sauvignon blanc in New Zealand (Bramley et al. 2011c, Trought and Bramley 2011), and in several cultivars in Australia (Bramley and Hamilton 2004).
There was considerable variation in most yield and grape composition metrics in both vintages for all four vineyards. Anthocyanins, phenols, and color intensity were the most variable in all four vineyards within and between vintages. A previous study found similar results, with anthocyanins and phenols being the most variable of grape composition metrics, with anthocyanins having CV values from 11.7 to 21.6% (Bramley 2005). Here, the CV for anthocyanins ranged from 10.4 to 19.2%. Among yield components, yield per vine was most variable, while berry weight was least variable. Yield per vine varied by five-fold, from the lowest in BP in 2009 to over 20-fold greater in the Lowrey vineyard in 2008. Substantial variation in yield was also found; up to 10-fold variation in a single vineyard has been reported (Bramley and Hamilton 2004).
Linear correlation
Numerous correlations were observed among soil, water status, yield, and berry composition variables. Soil moisture correlated inversely with berry weight and % silt and directly with color intensity, phenols, % clay, CEC, and soil pH. Many of these correlations were not intuitive: one would not expect increased soil moisture to result in decreased berry weight and concomitant increases in grape maturity and anthocyanins. However, all vines will not respond to soil water status in an identical fashion and others have found conflicting results (Koundouras et al. 2006). Mean Ψ (a.v.) correlated inversely with berry pH, TSS, anthocyanins, color intensity, and % clay, and directly with vine size and % sand. There was a relationship between vine water status and grape maturity: as leaf Ψ became more negative, the grapes accumulated more sugars and anthocyanins, consistent with previous reports for anthocyanins, but not soluble solids (Sivilotti et al. 2005). This occurred in soils with more clay and less sand. The higher clay soils may have had higher water content, but the vines were not accessing it and the higher sand soils were likely better drained, with less available water. Vine size was not highly correlated to any other variable, unlike in a previous study (Cortell et al. 2008). In 2009, there were fewer strong correlations between variables (Table 8). Mean soil moisture correlated inversely with anthocyanins and % silt, and directly with % clay, CEC, and soil pH. In both years, it was the soil texture, especially % clay, which largely determined soil moisture. Mean leaf Ψ (a.v.) correlated inversely with % clay content, and directly with % sand. Again, soil type was a major driver of the water status and terroir, consistent with previous findings (Seguin 1986). The % clay also correlated inversely with berry weight and % silt correlated directly with anthocyanins. Soil texture and water status, therefore, were both contributors to variability in grape composition.
Principal component analysis
The BP vineyard in 2008 was heavily associated with % clay, % OM, soil moisture, CEC, soil pH, and phenols, and was the most different from the other three blocks (Figure 7). RP1 and RP2 were similar to each other, without clear separation of water status zones within the vineyards. Lowrey vineyard was associated with mean leaf Ψ (a.v.), % sand, and vine size. Mean leaf Ψ correlated very negatively with TSS, berry pH, and total anthocyanins. As in 2008, BP in 2009 was very different from the other blocks and heavily associated with % clay (Figure 8). In 2009, the two RP1 and RP2 vineyards were more distinct from one another, with RP2 more heavily associated with berry TSS, anthocyanins, and hue. RP1 was associated with phenols and berry TA. Combining both vintages, berry composition variables (TSS, pH, anthocyanins, and phenols) were generally highly correlated to one another and negatively correlated with yield, cluster size, mean leaf Ψ, and % sand (Figure 9). In both years, but especially in 2008, it was a combination of vine water status and soil type that drove the composition of the grapes at harvest. Soil moisture was not a strong indicator of vine water status and vine size did not play a significant role in driving vineyard variability.
Geomatics and spatial relationships
This study demonstrated that geomatics tools can be used to characterize spatial variability in vineyard performance and to identify sources of that variation. This approach can be used by growers to understand and selectively manage spatial variation in vineyards to optimize whole vineyard performance.
Spatial variation in water status zones, the basis for intra-block divisions in this study, was stable between vintages (Figures 3 to 5). Use of k-means clustering established that assigning zones to high and low vine status categories based on median values was mostly accurate, although a small number of vines were mis-categorized. Vine water status showed few spatial patterns that were similar to yield and berry composition metrics; spatial relationships were apparent between leaf Ψ versus soil moisture and cluster weight, and inversely with anthocyanins (RP1, RP2, and BP), pH (RP1), TSS, and color (RP2). Vine size showed spatial distribution similar to other variables, even though there were no strong correlations between these variables. The 2008 map of vine size measurements for RP2 showed a region of higher vine size in the eastern half of the vineyard, where there was also higher soil moisture, lower anthocyanins, and higher TSS. This trend did not appear in the other vineyards. Lowrey vine size from 2008 had little similarity to the trends in the soil moisture, TSS, and anthocyanin maps. This was the only vineyard with vine size measurements in 2009 and again, there was no strong similarity to the maps of soil moisture, anthocyanins, or any other grape composition metric. Cortell and Kennedy (2006) mapped vigor zones in Pinot noir vineyards, and anthocyanin distribution in those same vineyards (Cortell et al. 2007a, 2007b, 2008), but they did not compare these maps directly. They found a strong relationship between vine vigor and anthocyanin composition, and clear spatial relationships between the vigor zones and anthocyanins.
Soil texture of the vineyard, another driver of variability, also matched spatially with other variables. In RP1 (Supplemental Figure 1), high and low % clay corresponded to zones of vine water status, yield components, and to some extent, grape composition. This same general relationship between % clay and other variables was present in the other vineyards as well.
The spatial variation in yield components and berry composition variables were, for the most part, consistent across the two seasons. Others also found stable and clear patterns in berry composition zones (Bramley 2005). These zones were related to those identified for yield (Bramley and Hamilton 2004), but in contrast to this study, there were similarities among spatial variations in berry weight versus grape composition variables. There were no distinct year-to-year patterns in grape composition or vine performance in a Riesling vineyard over a five-year period (Reynolds et al. 2007). The similarity of the weather in the two years of this study, although a complicating factor for observing effects of water status, may be partially responsible for the stability of these trends. The correlations described were generally not very strong, with r values >0.4 considered strong. The spatial variation apparent in maps of the variables indicates that the trends were not as strong as suspected. Further years of study are recommended to gain a better appreciation of trends in spatial variation over time.
Implications of weather
It is important to note that the two years of this study were abnormal in several ways. Both years were cooler and wetter than average (data not shown; Weather INnovations, Inc.). In contrast, 2007 was much hotter and drier than average and vine water stress incurred in 2007 may have impacted vine performance in 2008 as much as or more than excess moisture in 2008 and 2009. Rainfall events were more frequent in 2008, but the overall precipitation was higher in 2009. Daily max temperatures in 2008 were slightly above the long-term average, but min temperatures were below average, with a net effect of lower average temperatures. Max and min daily temperatures in 2009 were below average. The effect of the wet weather was apparent in the leaf Ψ measurements. In 2008, the lowest mean Ψ value was −0.94 MPa in the BP vineyard and in 2009, it was −0.95 MPa in both RP1 and Lowrey vineyards. Max mean Ψ was −0.60 MPa in 2008 and −0.59 MPa in 2009, both in the Lowrey vineyard. Stomatal opening, and consequently grapevine photosynthesis, are generally not affected by water stress until the midday Ψ is <−0.50 MPa (Hardie and Considine 1976). Even at their lowest leaf Ψ values, the vines in this study were likely not subjected to more than mild water stress.
Conclusions
All three vineyards were spatially variable in vine growth and performance, with berry anthocyanins and phenols being the most variable grape composition metrics. Yield and vine size were the two most variable attributes in all four vineyards in both years. Using water status as the basis for within-vineyard division into production lots, this study found no differences between water status zones in terms of vine size, berry composition, or soil properties. The weather in 2008 and 2009 complicated this study, as both years were wetter and cooler than average. This may have helped stabilize some spatial trends in vineyard performance, but also likely reduced any effect of water status on the vines. Leaf Ψ had some correlation to berry composition, but was not proven to be a driving factor in the spatial distribution of variability. The wetter than average years meant that the range of water status values was very narrow, with very little difference between high and low water status zones. Vine size was not a primary factor in driving vineyard variability and in the year that data were collected from all vineyards, it did not correlate well with maps of most other variables. Vine size maps bore some similarity to other variables, but not always. Spatial trends in variability of grape composition, soil moisture, and vine water status were generally stable from year to year. In addition, correlations between variables were confirmed in spatial distribution by qualitative comparison of maps. Although these results did not entirely prove our initial hypotheses that water status zones would correlate spatially with zones of significant berry composition variables, these observations are in themselves noteworthy, as they show that precision viticulture may not always be worthwhile in regions with substantial summer and autumn rainfall that may reduce zonal differences.
Acknowledgments
The authors thank Coyote’s Run Winery and Lowrey Five Rows Craft Winery, St. Davids, Ontario, for their cooperation. Funding by Ontario Centres of Excellence is likewise acknowledged.
Footnotes
Supplemental data is freely available with the online version of this article at www.ajevonline.org.
- Received June 2016.
- Revision received September 2016.
- Revision received October 2016.
- Accepted November 2016.
- Published online December 1969
- ©2017 by the American Society for Enology and Viticulture