Site selection

In April 2005, 10 Riesling vineyard sites were selected throughout the Niagara Peninsula in Ontario ( Supplementary Table S1 ¹). These sites were nonirrigated, commercial vineyards, and the vineyard blocks had heterogeneous soil types. Each site was also representative of each Vintners Quality Alliance (VQA) sub-appellation. Details concerning soil and vineyard characteristics and vineyard management can be found in  Supplementary Table S1 ¹ . All vineyards were balance pruned (Jordan et al. 1981) by retaining 20 nodes for each kilogram of cane prunings prior to each growing season. In each vineyard block, a grid-style sampling pattern was established with a sentinel vine at each grid intersection point. These sentinel vines (72–80 per vineyard block) were flagged for identification to be used for data collection. A Raven Invicta 115 GPS receiver (Raven Industries, Sioux Falls, SD, USA) with a built-in differential GPS correction receiver giving positions with an accuracy of 1–1.4 m was used in May 2005 to georeference each sentinel vine and to delineate the shape and size of each vineyard block.

Viticultural data collection

The SWC (%) and leaf ψ measurements were taken biweekly in each vineyard (every 10–14 d) from sentinel vines between the end of June and early September (beginning of fruit set to preharvest) as described in Willwerth and Reynolds (2020). Soil analyses, including pH, organic matter (OM) concentration, elemental concentration, cation-exchange capacity (CEC), and base saturation (BS), were performed on one soil sample collected from every fourth sentinel vine (Willwerth and Reynolds 2020). For each sentinel vine, data were collected annually at vine dormancy for weight of cane prunings as an estimate of vine vigor (vine size). Yield components (yield per vine, clusters per vine, cluster weight, berries per cluster, and berry weight) were either measured directly or calculated from measured variables during harvest each season. Fruit was sorted based on treatments and retained for winemaking. Clusters were counted from each sentinel vine, and samples of 100 berries were taken for determination of berry weight and standard fruit composition indices [Brix; titratable acidity (TA); pH], whereas samples of 250 berries were taken for monoterpene concentration analyses.

Fruit composition

Each 100-berry and 250-berry sample was weighed to determine the mean berry weight. The frozen berry samples were then heated in 250 mL beakers to an internal temperature of 80 °C in a Fisher Scientific Isotemp 228 water bath (Fisher Scientific, Ottawa, ON, Canada) to dissolve any precipitated tartaric acid. The heated berry samples were then cooled, juiced in a laboratory juicer (model 500; Omega Products Inc., Harrisburg, PA, USA), and a ≈35 mL portion was clarified using a IEC Centra CL2 centrifuge (International Equipment Co., Needham Heights, MA, USA). Total soluble solids (hereinafter Brix) were measured on the clarified berry juice samples using a temperature-compensated Abbé bench refractometer (model 10450; American Optical Corp., Buffalo, NY, USA). The pH was measured using an Accumet pH/ion meter AR50 (Fisher Scientific). TA was measured on 5 mL clarified samples using a Man-Tech PC-Titrate autotitrator (PC-1300-475; Man-Tech Associates Inc., Guelph, ON, Canada). Samples were titrated to a pH 8.2 endpoint with 0.1 mol L^-1 NaOH solution. Results were expressed as tartaric acid equivalents (g L^-1). Monoterpenes were analyzed for the 250-berry samples using the method developed by Dimitriadis and Williams (1984) as modified by Reynolds and Wardle (1989). The free volatile terpene (FVT) and potentially volatile terpene (PVT) concentrations were expressed as mg kg^-1. A large database was compiled annually on these sentinel vines for all yield and berry composition variables.

Statistical analysis

Linear and spatial correlations were determined between soil composition, soil texture, leaf ψ, SWC, and berry composition for all vintages. Through the use of XLSTAT (XLSTAT-pro version 2008.5.1), principal component analysis (PCA) was conducted to elucidate relationships among soil variables, soil and vine water status, and berry composition variables. Soil variables were used as supplementary variables for PCA. MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada) were used to construct geospatial maps of all variables. Spatial correlations were also determined using the statistical package provided by Vertical Mapper. These maps and the spatial correlations were used to examine spatial variation for selected variables in each season and their temporal stability and to compare apparent spatial relationships between correlated variables.

General comments

All results shown are from the 2005–2007 growing seasons. As in Willwerth and Reynolds (2020), three sites were omitted from this paper due to significant winter injury in the 2004–2005 winter that substantially decreased yields (Lambert, Reif, and Vailmont), and powdery mildew in 2006 (Vailmont). Meteorological data depicting temperature and rainfall events for each growing season are depicted in  Supplementary Fig. S1 ¹ . There were drought periods in all three growing seasons, but this was particularly the case in 2005 and the very hot and dry 2007 vintage. The 2006 growing season was marked by cooler and wetter periods later in the season, particularly in late August, September, and October. Results are subdivided into tables depicting linear correlations for each vineyard, geospatial maps depicting spatial variability in fruit composition variables for each site, and PCA results and interpretations. References to the companion manuscript (Willwerth and Reynolds 2020) are given for correlative relationships between leaf ψ and yield components.

Linear correlations

Correlations for each vineyard are depicted in separate Tables (Tables 1–7). Soil variables (e.g., % sand and clay, pH, OM, CEC, BS, elements), SWC, and leaf ψ are presented as putative independent variables and fruit composition entities as dependent variables. As most soil variables are relatively stable, particularly in soils with high clay content, whereas berry composition variables are inconsistent and subject to vicissitudes of annual weather patterns, it is not surprising that strong correlations were not abundant. Nonetheless, SWC and leaf ψ were frequently correlated to some berry composition variables.

Table 1.

Linear correlations between water status, soil, and berry composition variables for Glenlake Vineyards, Niagara-on-the-Lake, ON, Canada, 2005–2007.

Table 2.

Linear correlations between water status, soil, and berry composition variables for Chateau des Charmes, St. Davids, ON, Canada, 2005–2007.

Table 3.

Linear correlations between water status, soil, and berry composition variables for Paragon Vineyards, Jordan, ON, Canada, 2005–2007.

Table 4.

Linear correlations between water status, soil, and berry composition variables for Henry of Pelham Winery, St. Catharines, ON, Canada, 2005–2007.

Table 5.

Linear correlations between water status, soil, and berry composition variables for Myers Vineyard, Vineland, ON, Canada, 2005–2007.

Table 6.

Linear correlations between water status, soil, and berry composition variables for Flat Rock Cellars, Jordan, ON, Canada, 2005–2007.

Table 7.

Linear correlations between water status, soil, and berry composition variables for Cave Spring Cellars, Beamsville, ON, Canada, 2005–2007.

Spatial variability

At the GLK site, consistent zones of leaf ψ and SWC were found in 2005–2007 despite different weather conditions (Willwerth and Reynolds 2020). Maps for four sites (GLK, CDC, MYR, and CSC) are included in the main body of the paper (Figs. 1–11), while those for Brix and TA for the other three sites (PAR, HOP, and FLR) are included in  Supplementary Figs. S2– S4 ¹ . Spatial trends in Brix (Figs. 1A–1C) were inconsistent over the period of this study. There were some apparent spatial relationships with leaf ψ in some areas of the vineyard. The differences in Brix in 2005 might have been related to winter injury. TA (Figs. 1D–1F) and FVT (Figs. 2A–2C) showed evidence of temporal stability from 2005 to 2007. Strong evidence of spatial relationships was also found between leaf ψ and monoterpenes (Fig. 2), particularly in 2005. At CDC, many variables were temporally stable despite different weather conditions experienced in the 2005–2007 growing seasons. These included Brix (Figs. 3A–3C), TA (Figs. 3D–3F), and both FVT (Figs. 4A–4C) and PVT (Figs. 4D–4F). Leaf ψ was highly spatially related to both FVT and PVT in both 2005 and 2006 (Willwerth and Reynolds 2020). Leaf ψ and berry weight had strong spatial relationships with many fruit composition variables, including Brix, TA, and monoterpenes. Therefore, leaf ψ may have been impacting these variables indirectly through a higher skin-to-juice ratio caused by smaller berries. This site was one of the vineyards that demonstrated temporal consistency in terms of spatial variability.

Fig. 1.

Spatial patterns in soluble solids (Brix) (A–C) and titratable acidity (D–F). Glenlake Vineyard, Niagara-on-the-Lake, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 2.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Glenlake Vineyard, Niagara-on-the-Lake, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 3.

Spatial patterns in soluble solids (Brix) (A–C) and titratable acidity (D–F). Chateau des Charmes Vineyard, St Davids, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 4.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Chateau des Charmes Vineyard, St. Davids, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 5.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Paragon Vineyard, St. Catharines, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 6.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Henry of Pelham Vineyard, St. Catharines, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 7.

Spatial patterns in soluble solids (Brix) (A–C) and titratable acidity (D–F). Myers Vineyard, Vineland, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 8.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Myers Vineyard, Vineland, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 9.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Flat Rock Cellars, Jordan, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 10.

Spatial patterns in soluble solids (Brix) (A–C) and titratable acidity (D–F). Cave Spring Vineyard, Beamsville, ON, Canada. 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

Fig. 11.

Spatial patterns in free volatile terpenes (A–C) and potentially volatile terpenes (D–F). Cave Spring Vineyard, Beamsville, ON, Canada, 2005–2007. The figure was created through the use of MapInfo and Vertical Mapper (Northwood GeoScience, Ottawa, ON, Canada).

At the PAR site, spatial trends in Brix ( Supplementary Figs. S2A– S2C ¹ ) were relatively temporally stable from 2005 to 2007. In general, areas of higher Brix corresponded to areas of lower leaf ψ. TA ( Supplementary Figs. S2D– S2F ¹ ) had similar spatial trends to Brix but were inversely related. Similarly, leaf ψ (Willwerth and Reynolds 2020) seemed to have some relationships with berry composition in that areas of high leaf ψ also had higher TA. FVT (Figs. 5A–5C) and PVT (Figs. 5D–5F) both had some areas of temporal stability but were inconsistent in from year to year in general. HOP spatial trends in Brix ( Supplementary Figs. S3A– S3C ¹ ) were temporally stable from 2005 to 2007, and the values were also consistent from year to year, despite differences in growing seasons. TA ( Supplementary Figs. S3D– S3F ¹ ) was transient in nature: in 2005 and 2007 it was spatially similar, but in 2006 the trend was completely the opposite within the vineyard. FVT (Figs. 6A–6C) and PVT (Figs. 6D–6F) demonstrated some stability in terms of their spatial trends within the vineyard. Little variation was observed with respect to FVT, but there was more variation in PVT. Monoterpene concentrations were highest in 2006. Consistent leaf ψ zones were identified at the MYR site in 2005–2007. Spatial variability in fruit composition variables and spatial relationships among them were temporally stable (Figs. 7 and 8). There was low spatial variability in Brix, but spatial trends were temporally stable over the 3 yr of the study (Figs. 7A–7C). Brix demonstrated strong spatial relationships with SWC and leaf ψ (Willwerth and Reynolds 2020). Berry TA had areas of temporal stability, particularly in 2006–2007 (Figs. 7D–7F). The lowest TA occurred in 2007, likely due to the hot, dry growing conditions. Inconsistencies observed in 2005 were likely attributable to winter injury. Spatial relationships were observed in terms of TA and leaf ψ in that areas of higher leaf ψ had higher TA values. Brix and TA were inversely related spatially. Spatial trends of FVT (Figs. 8A–8C) and PVT (Figs. 8D–8F) were relatively stable over the 3-yr period, and there were some relationships with leaf ψ (Willwerth and Reynolds 2020).

Generally, areas of lower leaf ψ had higher FVT and PVT. Monoterpenes were highest in the 2006 vintage, particularly PVT. No clear spatial relationships existed between Brix and monoterpenes. At FLR, geospatial maps showed that many variables were temporally stable despite different weather conditions experienced in each growing season. Some spatial trends in terms of Brix ( Supplementary Figs. S4A– S4C ¹ ) were observed from 2005 to 2007 with few inconsistencies. Many of these can be related to areas of winter injury, particularly in the west end of the vineyard. TA ( Supplementary Figs. S4D– S4F ¹ ) was similar to Brix in terms of spatial variability and temporal stability. Areas of higher Brix had lower TA as expected. Values of Brix were relatively consistent over 3 yr but TA was highest in 2006. Spatial trends in FVT (Figs. 9A–9C) were inconsistent temporally, but spatial variability was not high. Stronger apparent spatial trends were observed with respect to PVT (Figs. 9D–9F). Trends were generally stable in PVT and were lowest in 2007. Contrary to other vineyards, CSC did not show temporal stability for many of the variables studied. Possible reasons for this are explained in Willwerth and Reynolds (2020). Only some areas of the vineyard were stable for fruit composition variables. Spatial trends in Brix (Figs. 10A–10C) were not temporally consistent from 2005 to 2007. TA (Figs. 10D–10F) spatial patterns were stable for 2005 and 2007, but in the wetter 2006 vintage the patterns were opposite. There were some areas showing temporal stability for FVT and PVT, whereas in some regions they were highest for both years (Figs. 11A–11F).

Principal components analysis

Through PCA, some relationships were observed at the GLK site (Figs. 12A–12C). In 2005, leaf ψ, SWC, and vine size were correlated. Leaf ψ was highly correlated with fruit composition. Lower leaf ψ values were related to higher Brix, terpenes, and pH in this vineyard for 2005. As expected, Brix, monoterpenes, and pH were all highly correlated. In 2006, SWC and leaf ψ were not correlated, but leaf ψ, Brix, vine size, and FVT were correlated. SWC was highly correlated with yield, TA, and % clay. In 2007, many variables were highly correlated. Leaf ψ and terpenes were highly correlated, as were OM and soil pH, while being inversely correlated to TA. SWC, Brix, pH, and % sand were also highly positively correlated to each other and inversely correlated to yield. At the CDC site, PCA suggested that leaf ψ was highly positively correlated with berry weight, Brix, and pH in 2005 and 2006 (Figs. 12D–12E). These same variables were also highly inversely correlated with FVT and PVT, vine size, and yield. The same trends were shown in 2007, except leaf ψ was noncorrelated with yield and pH (Fig. 12F). Therefore, low leaf ψ (i.e., water stress) seemed to have a positive impact on fruit composition resulting in higher terpene concentrations.

Fig. 12.

Principal component analysis of viticulture and soil variables, ON, Canada, 2005–2007. (A–C) Glenlake Vineyard; (D–F) Chateau des Charmes; (G–I) Myers Vineyard; (J–L) Cave Spring Vineyard. The figure was created through the use of XLSTAT (XLSTAT-pro version 2008.5.1) (XLSTAT-pro version 2008.5.1).

At the PAR site, PCA of 2005 data showed that SWC and leaf ψ were positively correlated with % clay, TA, and FVT, whereas berry weight, vine size, % sand, and Brix were inversely correlated ( Supplementary Fig. S5A ¹ ). In 2006, leaf ψ was positively correlated with vine size, berry weight, berry pH, TA, and FVT and inversely correlated with OM, P, and K. SWC was positively correlated with FVT, PVT, and Brix while being inversely correlated with yield ( Supplementary Fig. S5B ¹ ). In 2007, leaf ψ was correlated with Brix and pH and inversely correlated with OM, K, and P ( Supplementary Fig. S5C ¹ ). At the HOP site, PCA demonstrated that leaf ψ was positively correlated with TA and inversely correlated with % clay and pH in 2005 ( Supplementary Fig. S5C ¹ ). In 2006, leaf ψ was noncorrelated with Brix, pH, and TA, but correlated to monoterpenes ( Supplementary Fig. S5D ¹ ). In 2007, monoterpenes were inversely correlated with SWC, and Brix and TA were correlated with leaf ψ and % sand ( Supplementary Fig. S5E ¹ ). For the MYR vineyard, PCA showed that SWC, leaf ψ, vine size, berry weight, and Ca were correlated with TA in all three vintages (Figs. 12G–12I). These same variables were noncorrelated with Brix in the hotter and drier 2005 and 2007 vintages and inversely correlated with Brix in 2006. OM was associated with monoterpenes in 2005 and 2006, as was % sand in 2006. In general, yield was inversely related to Brix and terpenes. Brix and monoterpenes were generally highly correlated and inversely correlated with TA. At FLR in 2005, SWC and leaf ψ were noncorrelated; however, leaf ψ was positively correlated with PVT, K, and % sand while being inversely correlated with vine size and berry pH. SWC, however, was more correlated with TA and yield ( Supplementary Fig. S5G ¹ ). In 2006, leaf ψ and SWC were more closely related, showing a higher positive correlation, and leaf ψ was inversely correlated with yield, TA, and terpenes ( Supplementary Fig. S5H ¹ ). In 2007, leaf ψ, SWC, % sand, and OM were highly correlated with Brix and TA and inversely related to berry weight and terpenes, therefore indicating that lower leaf ψ and SWC resulted in wines with lower TA and higher flavor compounds ( Supplementary Fig. S5I ¹ ). At CSC in 2005, leaf ψ was highly correlated with PVT, % clay, CEC, and Ca while being noncorrelated with SWC (Fig. 12J). It was also inversely correlated with yield, berry weight, and vine size. SWC, however, was highly correlated with % sand, vine size, yield, TA, and P. FVT and Brix were correlated with % sand and soil pH. In 2006, leaf ψ and SWC were both highly correlated, as well as yield, vine size, yield, OM, and K (Fig. 12K). All of these variables were inversely correlated with Brix, TA, FVT, and PVT. In 2007, leaf ψ was directly correlated with Brix and pH and inversely correlated with monoterpenes (Fig. 12L).

Spatial trends and relationships within vineyard sites from 2005 to 2007 — Brix

Spatial variability in Brix was found in every vineyard site; however, in general, large variations in Brix did not exist. They varied in the range of 1.5–3 Brix within sites. This is agreement with research performed in Washington, where Brix varied in Concord vineyards (Davenport et al. 2001), and Australia (Bramley 2005). Brix levels were the highest in 2005 and lowest in 2006, which is indicative of those growing seasons. The 2006 vintage was cooler and wetter with more vegetative growth, which generally results in lower Brix levels. The 2005 growing season was very hot and dry, and the crop levels were very low due to bud damage caused by the winter of 2004–2005. The combination of these factors resulted in fruit with higher sugar levels than the other two vintages. This is consistent with Jackson and Lombard (1993) and Reynolds et al. (1994) with respect to Brix response to climate and crop level, respectively. There was, however, a large variation in Brix in some vineyards due to individual vine variation caused by cold injury, with levels as low as 16 Brix in some areas in 2005 and much higher Brix in subsequent years. The 2007 growing season was also hot and dry, but yields were higher due to lack of winter injury, resulting in higher than normal Brix levels in general. The variation was not as great in this particular vintage as it had ample sunshine, heat units, and extended dry periods.

Brix values showed evidence of temporal stability in seven vineyards over the 3-yr period. Through spatial and correlation analysis and PCA, there were often associations between SWC and Brix in many of the vineyard sites; however, they were not always consistent in terms of their relationship, or from year to year. In general, areas of lower SWC had higher Brix levels, but this was not always the case and some sites (CDC) consistently demonstrated the opposite effect. Leaf ψ was more consistent in its relationship with Brix. Leaf ψ was found through spatial analysis and PCA to have an apparent impact on Brix. In general, areas of vines with lower leaf ψ had higher Brix values than areas of higher leaf ψ. This also corresponds to lower SWC. This can be explained as being mainly due to the impact of water status on vegetative growth (i.e., weight of cane prunings), resulting in lower vine size. Some studies indicate that water stress can advance maturity (McCarthy and Coombe 1985) and increase Brix levels (Kliewer et al. 1983). These increases in Brix can be due to a concentration effect resulting from water loss, lower yields (Smart 1974) from water deficit, which reduces carbohydrate sinks in the vine, and better fruit exposure due to less vegetative growth (Smart 1974, 1985; Smart et al. 1990). Lower water status decreased shoot growth, and that water deficit accelerated sugar accumulation and malic acid breakdown (Koundouras et al. 2006). Brix values were also related to berry weight on a number of occasions. These may be a result of a concentration effect more than anything else. Some studies have shown that water stress can lead to higher final Brix levels in the fruit not through gains in sugar translocation but associated with reductions in berry weight (Hardie and Considine 1976). The findings of this study support that notion as well. In a few vineyards such as CDC, low leaf ψ vines had lower Brix. Under water deprivation, the decrease in photosynthesis and sugar export from the leaves can lead to this reduction in berry sugar accumulation (Quick et al. 1992; Rogiers et al. 2004). Esteban et al. (1999) demonstrated that higher water availability increased sugar levels due to increased photosynthetic activity or increased leaf area (Carbonneau et al. 1983). No consistent findings were found in terms of soil variables on Brix levels within sites. Soil texture played a minor role in its impact on berry Brix levels. Sand content was inversely related to Brix in a few vineyards as were Ca, soil pH, and OM. Soil K, P, and Mg also showed few relationships.

It therefore appears that variation in Brix was a result of SWC and leaf ψ and their impact on vine size and berry weight. The findings of this study are consistent with those of van Leeuwen et al. (2004), who studied the impact of different soil textures and grape maturation of Bordeaux varieties in France. Sandy soils had large berries with low sugars but high acidity. Clay soils resulted in berries with the highest sugars; these soil effects were influenced through vine water status. It was concluded that mineral nutrient uptake by the vine and (or) availability in the soil did not have a significant impact on fruit quality (van Leeuwen et al. 2004).

Titratable acidity

TA was highly variable within vineyard sites. It also was very transient, and spatial trends within vineyards often varied from year to year. As a result, spatial trends in TA were temporally stable in only some vineyards. In a number of cases, spatial trends in TA were consistent in two of three vintages. The hot and dry 2005 and 2007 vintages were consistent with each other; however, in other cases, 2006 and 2007 vintages demonstrated consistent spatial trends in TA. This was a result of individual vine variation due to cold injury suffered during the winter of 2004–2005. Some damaged areas of vineyards that were affected by winter damage had a number of secondary clusters on the vine that matured later than fruit formed from primary buds and thus had higher TA values. As individual vines became less variable (due to better vine health), more consistent spatial trends were observed. Many of the relationships involving TA were similar to Brix. SWC and leaf ψ had an impact in some vineyards as shown through spatial correlation and PCA. In general, areas of higher SWC and vines with higher leaf ψ also had higher TA, consistent with Koundouras et al. (1999). Similar to Brix, this is a result of more vegetative growth and poorer fruit development. A more vigorous canopy results in more shading of the fruit, which can decrease malic acid degradation in the fruit (Kliewer and Lider 1968). TA was normally inversely related to Brix, but in some cases it was noncorrelated. There were few consistent relationships between soil variables and TA. Soil texture seemed to impact TA in only a few cases, mostly as a result of higher vine size. TA was often correlated with % sand, larger vines, and higher yields. These findings are consistent with van Leeuwen et al. (2004), who demonstrated that sandy soils in Bordeaux had larger berries and higher acidity.

Monoterpenes

Monoterpenes are an important measurement of fruit composition as they play a major role in the varietal character of Riesling wines. As in the case of Brix and TA, monoterpenes varied spatially within vineyard sites. FVT concentrations were variable within vineyards; however, spatial trends within vineyards concerning FVT were not very temporally stable, partly due to how consistently low their levels were in most of the vineyard sites. Spatial trends in FVT were only temporally stable in three vineyards (CDC, GLK, and MYR). This lack of temporal stability is consistent with Reynolds et al. (2007), who found that monoterpene spatial distribution varied across different vintages in an Ontario Riesling vineyard. PVT also varied spatially within vineyards; however, spatial trends in PVT were more temporally stable in vineyards than FVT, as temporal stability was found in seven vineyards. Some vineyards showed a remarkable amount of temporal stability for monoterpenes. One of the resounding findings was that in every vineyard site, monoterpenes were the highest in the 2006 vintage. The 2005 and 2007 vintages were much hotter and drier than the 2006 vintage. These vintages had vines with lower leaf ψ. This suggests that periods of drought and resultant low leaf ψ have negative effects on monoterpene concentrations; however, regions within vineyard sites (e.g., MYR) with smaller berry weights seemed to have higher concentrations of FVT and PVT, reflecting the impact of skin-to-juice ratio. Therefore, it appears from these findings that either high (lack of any deficit) or very low leaf ψ can be detrimental to concentration in monoterpenes. Peyrot des Gachons et al. (2005) found similar results pertaining to vine water status and volatile thiols in Sauvignon blanc grapes and found that mild water deficits resulted in higher grape quality. In other studies, limited water availability also increased glucoconjugates of aroma compounds and increased wine quality (Koundouras et al. 2006). Reducing irrigation led to increased concentrations of monoterpenes (McCarthy and Coombe 1985), but others indicated that prolonged irrigation deficits reduced monoterpenes in berries (Reynolds and Wardle 1997; Reynolds et al. 2005). In the latter study, in Gewürztraminer grapevines, decreasing the duration of water stress increased FVT and PVT at harvest. Therefore, this research further supports the importance of vine water status and its impact on aroma potential in white wine cultivars and that mild water deficits appear to maximize aroma potential. No consistent relationships were shown between soil variables and monoterpene concentrations. It should be noted that in some instances, OM, P, and K were correlated with either FVT or PVT in multiple vineyards and vintages. Nutrient differences have been noted to possibly impact fruit composition. These results are not consistent enough to make any reasonable explanations; however, Reynolds et al. (2007) found similar relationships with PVT being correlated with P and K in Riesling, which is interesting and may require some elucidation through further research. Brix was generally correlated with monoterpene concentrations, which is an indication of general maturity. At times, TA was related to monoterpenes but no consistent trends were found.

Variables of fruit composition were neither as volatile nor as consistent as leaf ψ, vine size, and yield components, which is consistent with other precision viticulture studies that demonstrated that fruit composition at harvest was considerably less variable than yield (Bramley and Hamilton 2004; Bramley 2005). While there was evidence of noteworthy spatial relationships between leaf ψ and fruit composition, these relationships were not as strong as those between leaf ψ and other factors such as berry weight and vine size.

Through the use of geospatial technologies, many viticultural variables were found to be related to leaf ψ and vine size in many of the vineyard sites examined. As water deficits in vines can reduce shoot growth (Kliewer et al. 1983) and canopy density (Smart 1974), low water status may have altered the canopy characteristics of the vine, ultimately leading to better leaf and fruit exposure to sunlight and improved fruit composition. Furthermore, many of the relationships between vine size and leaf ψ demonstrate that water status is greatly influenced by water supply together with evaporative demand of the canopy. Thus, total leaf area and exposed leaf area measurements are relevant. Carbonneau (1999) has shown that ratio of exposed leaf area to yield is an important measurement in relation to vine vigor and that this can impact fruit quality due to sink competition. Similar to this study, Cortell et al. (2007) studied different vine size zones within vineyards and demonstrated that vine size impacted berry weight, Brix, and TA. They attributed variation between vine size zones to soil depth and water holding capacity. Greenspan and O’Donnell (2001) investigated the spatial variability within different vineyard blocks, and they were able to distinguish between high and low vine size zones and found that these management zones had different means of yield, Brix, and water status.

Variability is an important concept to understand for vineyard managers to produce consistent premium grapes in all wine-producing regions. In Ontario, there is an extreme amount of variability due to direct and indirect effects (i.e., cold injury) of this climate. Therefore, there is a need for necessary tools to study variation of vineyards. Variability in Brix levels is important to understand in situations where grapes are priced based on Brix such as in Ontario. Large variability can have negative economic consequences as one area of low Brix may reduce overall Brix levels, thereby impacting the potential quality and price of the grapes. Therefore, through precision viticulture a better understanding of grape maturity can be used to make wise harvesting decisions.