Abstract

Vineyards were planted in the arid region of northwest China to meet the local economic strategy while reducing agricultural water use. Sap flow, environmental variables, a plant characteristic (sapwood-to-leaf area ratio, As/Al) and a canopy characteristic (leaf area index, L) were measured in a vineyard in the region during the growing season of 2009, and hourly canopy stomatal conductance (Gsi) was estimated for individual vines to quantify the relationships between Gsi and these variables. After accounting for the effects of vapor pressure deficit (D) and solar radiation (Rs) on Gsi, much of the remaining variation of reference Gsi (GsiR) was driven by that of leaf-specific hydraulic conductivity, which in turn was driven by that of As/Al. After accounting for that effect on GsiR, appreciable temporal variation remained in the decline rate of GsiR with decreasing vineyard-averaged relative extractable soil water (θE). This variation was related to the differential decline ofθE near each monitored vine, decreasing faster between irrigation events near vines where L was greater, thus adding to the spatiotemporal variation of GsiR observed in the vineyard. We also found that the vines showed isohydric-like behavior whenθE was low, but switched to anisohydric-like behavior with increasingθE. ModeledθE and associated Gs of a canopy with even L (1.9 m2 m−2) were greater than that of the same average L but split between the lowest and highest L observed along sections of rows in the vineyard (1.2 and 2.6 m2 m−2) by 6 and 12%, respectively. Our results suggest that managing sectional L near the average, rather than allowing a wide variation, can reduce soil water depletion, maintaining Gs higher, thus potentially enhancing yield.

Introduction

Water resources are becoming increasingly limited in many regions worldwide due to both declining quality and availability, while demand is rapidly increasing with population size and per capita use rate (Vrörösmarty et al. 2000, Eckardt et al. 2009). Managers of water resources must balance allocation among various demands, but rarely meet any. For example, in the arid region of northwest China water shortages restrict regional economic development (Feng et al. 2000, Kang et al. 2008), yet light is amply available for fruit production. Recognizing these limitations and potentials, the strategy of local economic development encourages the establishment of vineyards, and wine production is becoming an important industry (Liu et al. 2006), thus increasing economic returns while consuming less water (383 mm in vineyards vs. 562 mm in wheat and maize fields over a growing season in this area; compare Zhang et al. 2011 with Ding et al. 2010).

In contrast to these annual crops, the canopy of vineyards is strongly coupled to the atmosphere (Lu et al. 2003), which promotes a higher potential water vapor transfer to the atmosphere (Jarvis and McNaughton 1986). Because in such well-coupled canopies both stomatal conductance and leaf area can exert great control over canopy-scale water loss, decreasing water consumption of vineyards compared with fields of annual crops must reflect a lower product of canopy leaf area index (L; see Table 1 for definitions of symbols and units) and average stomatal conductance. The primary purpose of this study was to assess the sensitivity of stomatal conductance in vineyards to environmental, vine and canopy variables. Such information would serve as an essential step in designing a method to optimize water use with respect to yield through tailored irrigation (Kim et al. 2008), and vine and vineyard architecture, and would help managers decide how to allocate limited water supply (Kang et al. 2004).

Table 1.

Symbols and their definitions.

Symbol Definition Unit 
Al Leaf area m2 
As Sapwood area at a height of 20 cm above the ground cm2 
C Universal gas constant m3 kPa K−1 kg−1 
D Vapor pressure deficit kPa 
D20 Diameter at a height of 20 cm above the ground cm 
d Characteristic leaf dimension 
El Transpiration per unit leaf area mmol m−2 s−1 
gs Leaf stomatal conductance mmol m−2 s−1 
gsi Leaf stomatal conductance of each individual mmol m−2 s−1 
Gbl Boundary layer conductance mmol m−2 s−1 
Gc Sap flow scaled total canopy conductance mmol m−2 s−1 
Gs Sap flow scaled canopy stomatal conductance mmol m−2 s−1 
Gsi Canopy stomatal conductance of each individual mmol m−2 s−1 
Gsi′ Modeled stomatal conductance of each individual mmol m−2 s−1 
GsR Reference canopy stomatal conductance mmol m−2 s−1 
GsiR Reference canopy stomatal conductance of each individual mmol m−2 s−1 
Kl Leaf-specific hydraulic conductance mmol m−2 s−1 MPa−1 
Ksh Thermal conductance constant of sap flow gauges W m−1 K−1 
L Leaf area index m2 m−2 
Li Leaf area index of 1 m to either side of each individual m2 m−2 
m Sensitivity of Gs to ln D mmol m−2 s−1 ln(kPa)−1 
RH Relative humidity – 
Rs Solar radiation W m−2 
Ta Air temperature °C 
W Wind speed m s−1 
Y0, a, b Parameters of relationship between Gsi with Rs andθE – 
θ Volumetric soil water content m3 m−3 
θi Volumetric soil water content near each individual m3 m−3 
θE Relative extractable soil water – 
θEi Relative extractable soil water near each individual – 
 Averaged volumetric soil water content over the vineyard – 
 Averaged relative extractable soil water over the vineyard – 
λa, λt Actual and theoretical ratio of m to GsR – 
ρw Density of water kg m−3 
ψmd, ψpd Midday and predawn leaf water potential MPa 
ψl, ψs Leaf and soil water potential MPa 
Symbol Definition Unit 
Al Leaf area m2 
As Sapwood area at a height of 20 cm above the ground cm2 
C Universal gas constant m3 kPa K−1 kg−1 
D Vapor pressure deficit kPa 
D20 Diameter at a height of 20 cm above the ground cm 
d Characteristic leaf dimension 
El Transpiration per unit leaf area mmol m−2 s−1 
gs Leaf stomatal conductance mmol m−2 s−1 
gsi Leaf stomatal conductance of each individual mmol m−2 s−1 
Gbl Boundary layer conductance mmol m−2 s−1 
Gc Sap flow scaled total canopy conductance mmol m−2 s−1 
Gs Sap flow scaled canopy stomatal conductance mmol m−2 s−1 
Gsi Canopy stomatal conductance of each individual mmol m−2 s−1 
Gsi′ Modeled stomatal conductance of each individual mmol m−2 s−1 
GsR Reference canopy stomatal conductance mmol m−2 s−1 
GsiR Reference canopy stomatal conductance of each individual mmol m−2 s−1 
Kl Leaf-specific hydraulic conductance mmol m−2 s−1 MPa−1 
Ksh Thermal conductance constant of sap flow gauges W m−1 K−1 
L Leaf area index m2 m−2 
Li Leaf area index of 1 m to either side of each individual m2 m−2 
m Sensitivity of Gs to ln D mmol m−2 s−1 ln(kPa)−1 
RH Relative humidity – 
Rs Solar radiation W m−2 
Ta Air temperature °C 
W Wind speed m s−1 
Y0, a, b Parameters of relationship between Gsi with Rs andθE – 
θ Volumetric soil water content m3 m−3 
θi Volumetric soil water content near each individual m3 m−3 
θE Relative extractable soil water – 
θEi Relative extractable soil water near each individual – 
 Averaged volumetric soil water content over the vineyard – 
 Averaged relative extractable soil water over the vineyard – 
λa, λt Actual and theoretical ratio of m to GsR – 
ρw Density of water kg m−3 
ψmd, ψpd Midday and predawn leaf water potential MPa 
ψl, ψs Leaf and soil water potential MPa 

Mean canopy stomatal conductance (Gs) is a key variable reflecting a crop's physiological response to changing environment in evapotranspiration models (e.g., Penman–Monteith model, Monteith and Unsworth 1990, Lebon et al. 2003), and can be used in assessing the local water balance of vineyards. Where the canopy is rough and its leaf area low, a condition that describes vineyards in general, the coupling of canopy and atmosphere is strong most of the time, and water loss is determined by the product of Gs and L (Jarvis and McNaughton 1986). Under non-limiting environmental conditions (sufficient light and soil moisture, low vapor pressure deficit, D, and optimal temperature), Gs is determined by leaf-specific hydraulic conductivity (Ward et al. 2008, Domec et al. 2009a, 2009b). For isohydric plants (i.e., those controlling leaf water loss such that xylem water potential remains above a tension that negatively affects tissue-specific hydraulic conductivity), leaf-specific hydraulic conductivity and, thus, Gs under standard environmental conditions should increase with the cross-sectional area of the stem (As) supplying water to a unit of transpiring leaf area (Al; Whitehead 1998).

Our first task, therefore, was to quantify and account for the commonly observed stomatal responses to incoming solar radiation (Rs) and D following the approach of Lu et al. (2003). This allowed us to focus on the effect of soil moisture (θ) and the ratio of sapwood area to leaf area (As/Al) on Gs (Schultz 2003, Soar et al. 2006, Williams and Baeza 2007). We hypothesized that Gs would increase to saturation withθ but linearly with As/Al (Domec and Gartner 2003, Ward et al. 2008). We accounted for the common responses of Gs to environmental variables based on the widely used empirical model formulated by Jarvis (1976), where the variation in Gs is partitioned among a series of multiplicative functions:  

(1)
formula
where Gs,max is maximum stomatal conductance when a plant is under optimal conditions, Rs is solar radiation, Ta is air temperature and D is vapor pressure deficit. Over the years, this model incorporating different functional relationships (Magnani et al. 1998, Granier et al. 2000, Uddling et al. 2004) has been successfully employed to describe the stomatal behavior of various cultivars of Vitis vinifera in the field at both the leaf (Winkel and Rambal 1990) and canopy (Lu et al. 2003) scales. However, parameter values of such models must be adjusted seasonally and among vineyards (Lu et al. 2003). This may reflect the variability of hydraulic and canopy attributes, suggesting that these attributes must be explicitly incorporated into the model when employed at the vineyard scale, over long periods, or both.

Many studies have quantified the magnitude and dynamics of vineyard canopy water use (Trambouze and Voltz 2001, Williams et al. 2003, Yunusa et al. 2004, Zhang et al. 2007, 2008, 2011, Li et al. 2009). But few studies in vineyards have assessed the response of Gs to both biological and environmental variables, and even fewer have investigated the spatiotemporal variation of Gs at a scale smaller than a vineyard. These few studies are more common in forest settings (Oren et al. 2001, Chang et al. 2006, Ewers et al. 2007, Herbst et al. 2008) and rely on measurements of whole-plant water uptake based on one of several sap flow techniques (Grainer 1987, Grime et al. 1995, Čermák et al. 2004, Lu et al. 2004), which can be used to estimate a continuous record of Gs by inverting the Penman–Monteith model (Köstner et al. 1992, Wullschleger et al. 2000, Lu et al. 2003).

In most physiological studies, soil moisture is measured at one or a few locations. Yet soil moisture is highly variable in space (Katul et al. 1997, Hupet and Vanclooster 2004, Jost et al. 2005), and although only a few sensors are required to trace the temporal changes inθ reasonably well, this information does not capture the spatiotemporal variability of water supply and its changes over drying cycles experienced by individual plants. In irrigated systems, water availability may be homogenized, reaching a maximum after irrigation, yet the rate of decrease in water availability, i.e., the depletion rate ofθ over drying cycles, may vary among patches depending on the variation of the amount of transpiring foliage. As a result, sap flux density may decrease with L, or the sum of As per unit of ground area, as soil drying progresses (Oren et al. 1998). We thus hypothesized that, after accounting for the atmospheric effects of Rs and D and the hydraulic effects of As/Al, Gs of vines positioned in row sections of higher L will decrease faster, and will be lower at a given vineyard-averagedθ than vines growing where canopy L is lower.

Sap flow and environmental factors were measured in a vineyard in Shiyang river basin in northwest China during the growing season of 2009. Structural attributes that may influence Gs, such as As/Al and L, were also measured. In a previous paper, grapevine sap flow measurement has been validated both by balancing the water budget using the upper 1 m of the soil, and by comparing the evapotranspiration component of the water balance with estimates based on the Bowen ratio energy balance (BREB) method (Zhang et al. 2011). In this study, we focused on Gs and quantified the effects of environmental conditions (Rs, D andθ), a vine-scale hydraulic characteristic (As/Al) and a vineyard scale canopy characteristic (L) on Gs, and combined with L, on water use. We note that the latter three attributes can be altered through irrigation, by pruning and changing vine density, thus facilitating optimization of yield-based water use efficiency.

Materials and method

Setting

The experiment was conducted in a vineyard of the Experimental Station for Water-Saving in Agriculture and Ecology (37°52′20″N; 102°50′50″E, 1585 m asl) of China Agricultural University in the Shiyang River basin, northwest China. The region is located in a continental temperate zone with an arid climate and a mean annual precipitation of 165 mm. The soil is a light sandy loam with a mean bulk density of 1.47 g cm−3 to a depth of 1.0 m and there is an impermeable stratum at a depth of 1.2 m according to the soil profile analysis.

The vineyard was established in 1999 over an area of 400 ha. Vitis vinifera L., cv. Merlot (own-rooted) were planted at a spacing of 1 m between vines and 2.7 m between rows oriented east–west (Figure 1). The vines were trained to have two trunks at the surface of the ground. In this study, we consider an individual to be one of the trunks. Shoots were trained to a vertical plane by three wires supported by a 1.5-m-high trellis system as a vertical shoot positional training system. The vineyard was furrow-irrigated with a 0.3-m-deep trapeziform furrow on the south side of each vine row, with surface and bottom widths of 1.0 and 0.9 m, respectively. The vineyard was irrigated five times during the 2009 growing season: 23 April, 14 May, 27 June, 24 July and 22 August. The irrigation volume in each event was 60 mm of water expressed as average for the entire vineyard area, but inter-furrow strips remained dry, indicating that the irrigation intensity was higher where the vine roots were growing.

Figure 1.

Layout of all the measurements.

Figure 1.

Layout of all the measurements.

Environmental monitoring

Precipitation (P, Rain Gauge Smart Sensor, model No. S-RGB-M006) and Rs (Silicon Pyranometer Sensor, model No. S-LIB-M003) were measured with a Hobo weather station (Onset Computer Corp., Bourne, MA, USA) located in a reference site ∼1 km to the west of the vineyard. Air temperature (Ta), relative humidity (RH) and wind speed (W) were measured by a BREB system (Campbell Scientific Inc., Logan, UT, USA) positioned at a height of 2.3 m in the vineyard and ∼30 m away from the sap flow site. D was calculated from Ta and RH following the functions given by Campbell and Norman (1998).

About 90% of fine root biomass and surface area (diameter ≤ 2 mm) were distributed in 0.1–0.8 m (Y. Zhang, unpublished data). Moreover, both variables decreased linearly with depth from peaks at the 0.3–0.4 m layer, dropping to ∼2% of total amounts in the 0.9–1 m layer, or half of the layer above. This is not surprising considering that soil profile analysis showed that sandy-loam soil uniformly distributes down to 1.2 m, where it meets an impermeable layer. Thus, soil volumetric water content (θ) was determined with a portable time-domain frequency probe measuring in fixed-position tubes (Diviner 2000, Sentek Pty. Ltd, Stepney, SA, Australia) down to 1 m. Twelve access tubes were installed at 10 m intervals in the irrigation ditches along the rows of vines (Figure 1). Soil moisture measurements were conducted weekly at 0.1 m intervals and were calibrated based on the gravimetric sampling technique.

To compare across sites and to normalize the influence of soil texture onθ, drought intensity is best quantified in the form of relative extractable soil water (θE, dimensionless), determined as  

(2)
formula
whereθt is the actualθ measured at time twm is the soil moisture of wilting point (0.11 m3 m−3) determined from soil water characteristic curves (CR 22G high-speed centrifuge, Hitachi, Tokyo, Japan) andθfc is the field capacity, which was determined as 0.31 m3 m−3 gravimetrically based on soil and agricultural chemistry analysis (Bao 2000). We averagedθE of the 12 tubes to estimate the average relative extractable soil water of the vineyard (graphic). Linear interpolation was applied between consecutive irrigations to determine graphic at each day of the growing season. graphic of days with rainfall was recalculated by adding the amount of rainfall.

To determine the specificθE of each monitored vine, we extracted soil samples with a ∼2-cm-diameter auger near each individual at the beginning and end of two drying cycles following the irrigations on 24 July and 22 August. Two sampling points were selected within 0.3 m of the monitored vine and 10 vertical samples (0.1–1 m deep) were taken at each point. The water content at the end of the drying cycle was compared with that at the beginning of the cycle. The relative decrease ofθ was then related to the local L.

Sap flow measurement

The heat balance method (Sakuratani 1981) was used to measure sap flow (Dynagage Flow32-1K system, Dynamax, Houston, TX, USA). Sap flow was monitored on 8 trunks from 28 April to 12 August 2009, and on 16 trunks from 13 August to 5 October 2009. Vines were chosen randomly along two sets of three rows within a 30-m-diameter distance (the length of the longest cable) from the data logger (Figure 1). The gauges were installed at a height of >0.2 m above the ground surface. Gauge output (sap flow volume through the entire stem, g h−1) was monitored every 60 s and recorded as 15-min means with a CR1000 data logger (Campbell Scientific, Logan, UT, USA). Since irrigations have resulted in overflooding the soil, sensors were disconnected before irrigation to avoid damage, and reinstalled within 3 days after irrigation. Further detail on the methods, theoretical background and installation procedure can be found in Trambouze and Voltz (2001).

To minimize thermal damage to vines (e.g., Stöhr and Lösch 2004), a night-time power down model was implemented in all but the nights used to determine the thermal conductance constant (Ksh). Ksh of each gauge was determined at every installation from the heat balance function on 2–3 nights of full power, assuming that no sap flow occurs before sunrise. Days with rains and low wind speeds were selected as likely to provide conditions meeting this assumption. In addition, full power at night was applied on 15 consecutive days to calculate the potential error on Ksh determination introduced by the assumption of no night-time sap flow. The resulting potential errors of flow estimation could be determined as the difference between two calculations, one based on the lowest Ksh of each night, and the other based on the lowest Ksh during the entire period, corresponding to the night with the lowest D and wind speed.

Canopy stomatal conductance calculation

Canopy stomatal conductance (Gs) was estimated from sap-flow-scaled total canopy conductance (Gc) after adjusting for boundary layer conductance (Gbl):  

(3)
formula

G c was calculated from hourly values of sap flow rate normalized by leaf area (El), assuming no storage and, thus, no lags between sap flow and total transpiration, using simplification of the inversion Penman–Monteith model (Köstner et al. 1992):  

(4)
formula
where Ta is air temperature in °C, C is the universal gas constant adjusted for water vapor and ρw is the density of water at Ta. Determination of leaf area for every individual is shown in the next section. Only conditions of D ≥ 0.6 kPa (∼66% of the data) were used to ensure that errors in estimates of Gc remain <10% (Ewers and Oren 2000). Gbl was determined from wind speed (W) and characteristic leaf dimension (d = 0.068 m) as in Campbell and Norman (1998):  
(5)
formula

Relating Gc to Gs produced a slope of 0.85 (data not shown), consistent with other studies showing high Gbl in vineyards (Lu et al. 2003).

Sapwood and leaf area determination

Given the anatomy of vines and the small stem diameters in the vineyard, we considered the entire cross-sectional area inside the thin bark to be the hydro-active xylem (sapwood) area (Braun and Schmid 1999). The diameter was measured at 0.2 m above the soil surface (D20).

During the period of leaf and shoot expansion from 28 April to 30 June, L was determined every 3–7 days. Each time, the number of shoots on each vine trunk was counted (N). Three shoots of each trunk were randomly chosen and all leaves on each shoot were classified into three categories by size, and counted (ni, i = 1–3). Three leaves were collected to determine the average individual area of leaves in each category (aii = 1–3). The leaf area of a single vine trunk (Al) was estimated based on the following:  

(6)
formula

After leaf expansion was completed (30 June), the dynamics of L (but not the absolute values) were quantified by a Sunscan canopy analyzer system (Delta-T Devices, Cambridge, UK). Canopy leaf area observations remained unchanged from 1 July to 10 September (P = 0.88 for difference among sampling times), and decreased to 75% of the maximum value on 27 September, the last L measurement date. Based on this and direct observations, we defined the period from 1 July to 10 September as full canopy with constant leaf area during which current year shoots were fully developed and no significant leaf fall occurred.

The growing-season dynamics of L were used to obtain two variables: (i) Al of each individual—necessary in calculations of mean stomatal conductance from sap flow, and (ii) L of each 2-m-long section along the row containing a monitored vine (Li)—necessary to test the hypothesis of faster decline ofθ with higher Li. To obtain these variables, Al following the leaf expansion period was estimated once (in August) by establishing a relationship between leaf area and shoot length. Leaf area was obtained by direct measurement of leaf area (AM300 portable leaf area meter, ADC Ltd, Herts, UK) from a sample of shoots (45 shoots were chosen to cover the range of shoot length from 0.05 to 1.10 m). The resulting relationship (Al = 20.49 × shoot length (cm) − 69.23, R2 = 0.96, P < 0.001) was employed to estimate Al of each vine and of each 2-m section near sap flow vines.

To estimate Al of each vine over the latter part of the season, Al was converted to individual stem L = Al/(2.7 m × 0.5 m ground area per stem) = Al/1.35 m2. The value was divided by the optical estimate of individual L, generating a conversion factor (0.61 ± 0.21) from the optical measurements, thus allowing conversion of the dynamics of each individual over the remaining part of the season to actual Al. To estimate values of Li, Al estimated allometrically over the 2-m sections was converted using Al/(2.7 m × 2 m ground area) = Al/5.4 m2. This, again, was divided by the section's Li, generating another, similar yet not identical, conversion factor (0.62 ± 0.18) to be applied to convert the seasonal dynamics of the optical measurements to obtain absolute values of Li. In days without optical measurements, Al and Li were estimated by a second-order polynomial regression with the number of days after bud break (R2 = 0.98, P < 0.001; Figure 2d).

Figure 2.

Seasonal dynamics of environmental factors and canopy characteristics: (a) precipitation (P) and irrigation (I), (b) solar radiation (Rs), (c) vapor pressure deficit (D), (d) leaf area of half-vines and (e–h) soil volumetric water content (θ) over a depth of 1.0 m and for depths of 0–0.29 m, 0.30–0.79 m and 0.80–1.00 m. The dashed lines in (d) are days when leaf area stopped increasing and chlorosis was noticeable in the vineyard, respectively. Error bars in (d) are standard errors of N = 8 individuals, and those in (e–h) are standard errors of N = 12 sampling positions (see Figure 1); dashed lines in (e–h) indicate the irrigation events.

Figure 2.

Seasonal dynamics of environmental factors and canopy characteristics: (a) precipitation (P) and irrigation (I), (b) solar radiation (Rs), (c) vapor pressure deficit (D), (d) leaf area of half-vines and (e–h) soil volumetric water content (θ) over a depth of 1.0 m and for depths of 0–0.29 m, 0.30–0.79 m and 0.80–1.00 m. The dashed lines in (d) are days when leaf area stopped increasing and chlorosis was noticeable in the vineyard, respectively. Error bars in (d) are standard errors of N = 8 individuals, and those in (e–h) are standard errors of N = 12 sampling positions (see Figure 1); dashed lines in (e–h) indicate the irrigation events.

Gas-exchange measurements

Although we used sap-flux-scaled conductance in most of our analyses, due to potential autocorrelation between Gs calculated as proportional to 1/D and D itself, we also assessed how well the crown-averaged Gs reflects leaf-level gs from independently obtained gas-exchange measurements. Leaf scale gas-exchange parameters were measured on 27 August with a Li-6400 portable photosynthesis system (Li-Cor Inc., Lincoln, NE, USA) on four individuals equipped with sap flow sensors. Eight leaves (two leaves from each of four canopy positions: on the south and north side of the vine from upper and lower canopy) of each individual were selected and measured once every 2 h from 8:00 to 18:00. The leaf chamber was held to keep the leaves in their natural positions. Stomatal conductance (gs) was averaged among the eight observations on each individual to represent the mean value of that individual at the measuring time. The value, implicitly assuming that leaf area was equally represented by the four positions, was compared with the corresponding Gsi calculated from the sap flux at the same time.

Water potential measurements and total hydraulic conductivity

Predawn (ψpd) and midday (ψmd) leaf water potentials were measured with a pressure chamber (SKPM1400, Skye Ltd, Llandrindod Wells, UK). Immediately after leaves were harvested, they were enclosed in a humid plastic zip lock bag (with moist filter paper), stored in a cooler and sent to the lab after all the samples were collected. Samples were measured immediately after arrival at the lab. It took ∼15 min from the start of collection to the start of measurements. Leaves for ψmd were picked at the radiation noon rather than the time of maximum D. The diurnal D pattern of most of the days during the study period had a relatively stable maximum plateau rather than a clear maximum, with high values stretching from the Rs peak for several hours. Thus, we collected leaves at the time in which light was not limiting and D was stable. Four campaigns were conducted (21 and 28 July, 17 and 27 August) before and after two of the irrigations (24 July and 22 August). The first two campaigns focused on four individuals, eight leaves (two leaves from each of the four canopy positions: on the south and north side and upper and lower canopy) of each individual. A two-way analysis of variance found no differences among different canopy positions for the predawn leaf water potentials (P = 0.53). For the midday values, upper canopy leaves had lower values than those from the lower positions (P = 0.03), but there was no directional effect (P = 0.45). Thus, during the last two campaigns, three leaves from the upper canopy were collected from all 16 individuals to obtain the lowest midday leaf water potential. The leaf-specific hydraulic conductance (Kl) was calculated as  

(7)
formula
where ψs is the soil water potential and ψl is the leaf water potential. ψs and ψl were substituted by the averaged ψpb and ψmd, respectively, in our study. The highest values of El were used because at this time light is saturating and leaves are likely to have transpired all stored water so that steady-state water flow conditions are likely to exist.

Leaf chlorosis estimation

To assess how well we can account for the temporal and spatial variation of Gsi, we utilized data obtained from an independent group of vine stems (Figure 1) during the end of the growing season. In this analysis, we had to account for an additional process: although leaf fall had not begun until day of year (DOY) 270, leaf chlorosis—indicating that senescence had already begun—was observed since DOY 244 (the dashed line in Figure 2d). Not accounting for this would result in overestimation of functional leaf area and underestimation of Gs (Steduto and Hsiao 1998). To convert total Li and As/Al to functional values, we predicted Gsi values over the entire study for the group of eight vines previously analyzed and then interpreted the end of season difference between modeled and actual values to reflect a progressive increase in non-functioning leaf area.

Data analysis

Analysis of the relationship between individual vine Gs (Gsi) and D, Rs and graphic was performed on hourly data partitioned into categories of solar radiation and soil water. To assess whether there are interaction effects of soil water and solar radiation on Gs, we first analyzed the influence of each factor when the other was at a relatively high level. Data during periods of graphic > 0.180 m3 m−3 (i.e., graphic > 0.29) were partitioned into six Rs categories (one <150 W m−2, one from 150 to 220 W m−2, three from 220 to 760 W m−2 at 180 W m−2 intervals and one >760 W m−2), and data during periods of Rs > 220 W m−2 were partitioned into five graphic categories (one <0.21, three classes from 0.21 to 0.47 of ∼0.09 intervals and one >0.47). With eight individuals involved, this resulted in 48 and 40 subsets for Rs and graphic analysis, respectively. We then focused on the conditions in which soil moisture was lower, for the range of < 0.29, and analyzed the response to Rs. If there is no interaction effect on Gsi between graphic and Rs, the response to graphic; would be lower by the same fraction along the entire range of Rs. There were insufficient data to properly define the effect of D, Rs and graphic on Gsi of the eight vines on which monitoring began later on 13 August, and these were left for independent validation of the capacity of the approach to explain the variation in observed Gs.

The boundary line for each subset of data was determined based on Schäfer et al. (2000) and Kim et al. (2008) as follows: within each Rs × graphic category, (i) individual vine Gs data (Gsi) were further partitioned to at least five D intervals, (ii) the mean and standard deviation of the Gs data within each interval were calculated, (iii) outliers were removed (P < 0.05, Dixon's test), (iv) data above the mean plus one standard deviation of Gs were selected, and (v) the selected data for each D interval with more than three remaining Gsi values were averaged. For each individual within Rs × graphic class, the mean Gsi values of all D intervals obtained in step (v) were related to the corresponding average D based on (Oren et al. 1999)  

(8)
formula
where GsiR is Gsi at D = 1 kPa (hereafter, reference canopy stomatal conductance) and m is the rate of reduction of Gsi per unit ln(D), reflecting the sensitivity of Gs to D. The theoretical ratio of m to GsiR can be calculated from the ratio of Gbl to maximum Gsi and the range of D following Kim et al. (2008). In the calculations, the effects of the path length and hydrostatic pressure were ignored, assuming that both are negligible in such short plants.

We then analyzed the influence of Rs and graphic on GsiR to quantify the remaining, unexplained variation. The response of GsiR to Rs and graphic was expressed as an exponential rise to a maximum (Jarvis 1976). Statistical analyses and regressions were conducted in SPSS 13.0 and SigmaPlot 10.0 (SPSS Inc., Chicago, IL, USA).

Results

Environmental variables

The total precipitation in the research period (28 April to 5 October) was 119 mm, occurring mostly in small events distributed over the season (Figure 2a). Daily average Rs and D, with similar seasonal dynamic courses, ranged from 30.4 to 324.6 W m−2 and from 0.16 to 2.83 kPa, respectively (Figure 2b and c).

There was no difference (P = 0.714) betweenθ averaged from the population of the four sensors near the vines monitored for sap flow and that of the population of eight sensors farther in the vineyard. graphic of the upper 1 m fluctuated from 0.147 to 0.253 m3 m−3, reaching a similar value after each of the four irrigations (Figure 2e). We estimated the coefficient of variability (CV = standard deviation/mean) ofθ following irrigations and before much depletion occurred, indicating the variability in water-holding capacity, to quantify the spatial variability of the soil physical characteristics. Mean soil moistures after each irrigation event ranged from 0.238 to 0.253 m3 m−3. The standard deviation of these mean values of the 12 sampling locations ranged from 0.021 to 0.026 m3 m−3, resulting in CV ranging from 8 to 11%.

Viewed by horizons,θ variation in upper soil layers was much greater than in deeper layers (see Figure 2f–h), with values fluctuating between 0.110 and 0.317 m3 m−3 in the upper 0.1 m, but remaining between 0.192 and 0.226 m3 m−3 at the bottom 0.1 m because deeper horizons did not show much water input signal (not shown).

Leaf area of single stems and vineyard sections

The leaf area of individual stems increased to DOY 181 (the dashed line in Figure 2d), showing slight fluctuations until the end of the season when leaf fall began before the termination of the study (Figure 2d). The switch from the method used during leaf expansion to the more common method used during the stable leaf area period did not manifest in an obvious discrepancy between the two periods, serving as a validation of the method used to follow the early dynamics. The stem diameter of the monitored vines ranged from 15.4 to 33.8 mm. During the stable leaf area period, sapwood area and leaf area varied among vine stems, ranging from 1.87 to 8.98 cm2 and from 0.74 to 4.80 m2, and with CVs of 24 and 30%, respectively. Variation of Al among individual stems translated to variation of Li in the 2 m row sections encompassing sap flow measurement stems, ranging from 1.2 to 2.6 m2 m−2, with a CV of 21%.

Potential errors caused by Ksh determination

During the 15-day-long investigation, one day, characterized by higher night-time maximum wind speed (3.7 vs. 1.4 m s−1) and D (1.1 vs. 0.5 kPa) than the averaged value of the rest of the period, showed that not accounting for night-time flux on the determination of Ksh following such unusual nights may result in 9.5 ± 2.1% error of daily flux among individuals. Maximum daily error of any individual over the rest of the 15-day period was <4%, averaging 2.4% over all days and individuals (Table 2). Wind speed and D conditions during the majority of the 15-day investigation were similar to those during nights on which we determined Ksh: wind speed during the calibration nights remained below 2 m s−1 and D was below 0.6 kPa. Thus, daily underestimates of flux caused by biased determination of Ksh, due to underestimation of night-time sap flow, were <5%.

Table 2.

Minimum (Min), maximum (Max), mean and standard deviation (SD) of errors caused by Ksh determination with no night-time sap flow assumption for the 15 days with full power supply at night. These parameters of night-time (23–4 h) mean and maximum D and wind speed of these days were also shown in different columns.

 Errors SD among individuals Dmean Dmax Wmean Wmax 
Min 0.77 0.54 0.19 0.21 0.29 0.29 
Max 9.54 2.66 0.91 1.07 2.81 3.70 
Mean 2.91 –_ 0.43 0.53 1.13 1.65 
SD 2.09 –_ 0.20 0.24 0.72 1.06 
 Errors SD among individuals Dmean Dmax Wmean Wmax 
Min 0.77 0.54 0.19 0.21 0.29 0.29 
Max 9.54 2.66 0.91 1.07 2.81 3.70 
Mean 2.91 –_ 0.43 0.53 1.13 1.65 
SD 2.09 –_ 0.20 0.24 0.72 1.06 

Conductance at the leaf and crown scales

Because the analyses designed to separate the contributions of environmental variables, hydraulic properties and competing L on stomatal conductance were conducted based on Gs estimated from sap flow and D, we first tested whether these estimates were similar to generally accepted values obtained with a gas-exchange system. A short, intensive campaign demonstrated that gs under prevailing conditions was 39% lower (P < 0.001) in the lower portion of the northern side of the rows than in the more illuminated lower portion of the southern side and the upper portion of the canopy of both sides (inset in Figure 3). Averaging the measurements obtained from the leaves of each stem (thus assuming each of the four positions represented a quarter of the stem's leaf area) produced similar values to Gsi (Figure 3), increasing confidence in conclusions regarding controls of stomatal conductance based on subsequent analyses of Gs.

Figure 3.

Comparison of canopy conductance calculated from sap flow (Gsi) and average stomatal conductance from leaf gas-exchange measurements of four canopy positions (gsi). Different symbols are different individuals, and each symbol is the data from 1 h. The dashed line is the 1:1 line and the solid line is the linear regression of the two variables. Inset figure shows gs of each of the four canopy positions (S, southern; N, northern; U, upper; B, bottom).

Figure 3.

Comparison of canopy conductance calculated from sap flow (Gsi) and average stomatal conductance from leaf gas-exchange measurements of four canopy positions (gsi). Different symbols are different individuals, and each symbol is the data from 1 h. The dashed line is the 1:1 line and the solid line is the linear regression of the two variables. Inset figure shows gs of each of the four canopy positions (S, southern; N, northern; U, upper; B, bottom).

The variation in Gsi was influenced by D, Rs and graphic. Gsi of all individuals decreased as D increased in the range of 0.6–4.8 kPa at each Rs and graphic category, for Rs > 150 W m−2 (see the example in Figure 4). The GsiD relationship (Eq. (10)) of all categories with Rs > 150 W m−2 was significant at P = 0.05 with R2 > 0.80. For Rs < 150 W m−2, Gsi was insensitive to D, but increased linearly with Rs, as we demonstrate later.

Figure 4.

An example of the relationship between canopy stomatal conductance calculated from sap flow (Gsi) and vapor pressure deficit (D) for different light levels. Small open circles are data with solar radiation (Rs) >150 W m−2; small solid circles are data with Rs < 150 W m−2. Large symbols are the boundary data for different light levels; boundary line fit curves are shown with bold solid lines for light classes of 150 ≤ Rs < 220 W m−2, 220 ≤ Rs < 400 W m−2 and Rs ≥ 400 W m−2, respectively. The fit curves of the highest three light classes (400 ≤ Rs < 580 W m−2, 580 ≤ Rs < 760 W m−2 and Rs < 760 W m−2 shown as solid triangles, open triangles and solid squares, respectively) were nearly indistinguishable, and for clarity only one curve was fitted and shown. Gray circles are data with D in the range of 0.9–1.1 kPa and Rs ≥ 150 W m−2.

Figure 4.

An example of the relationship between canopy stomatal conductance calculated from sap flow (Gsi) and vapor pressure deficit (D) for different light levels. Small open circles are data with solar radiation (Rs) >150 W m−2; small solid circles are data with Rs < 150 W m−2. Large symbols are the boundary data for different light levels; boundary line fit curves are shown with bold solid lines for light classes of 150 ≤ Rs < 220 W m−2, 220 ≤ Rs < 400 W m−2 and Rs ≥ 400 W m−2, respectively. The fit curves of the highest three light classes (400 ≤ Rs < 580 W m−2, 580 ≤ Rs < 760 W m−2 and Rs < 760 W m−2 shown as solid triangles, open triangles and solid squares, respectively) were nearly indistinguishable, and for clarity only one curve was fitted and shown. Gray circles are data with D in the range of 0.9–1.1 kPa and Rs ≥ 150 W m−2.

The sensitivity of Gsi to D, the parameter m in Eq. (8), was proportional to GsiR (P < 0.001). The intercept was not significantly different from zero (P = 0.888) and the slope of the fit line through the origin was 0.55 when pooled for all individuals andθE, and Rs categories >150 W m−2 (Figure 5a and b). The actual ratio of m to GsiRa) was lower (P < 0.001) than the theoretical ratio (λt = 0.60, Oren et al. 1999, Kim et al. 2008). The ratio λat of different Rs categories ranged from 0.80 to 1.09, averaging 0.91 for all individuals, and was insensitive to Rs (Figure 5c). In contrast, λat averaged 0.98 at low soil moisture, but decreased with graphic as values increased above 0.35 (Figure 5d), thus explaining the lower than expected ratio seen with respect to Rs, for which all analyses relied on high soil moisture conditions. The pattern of λat relative to graphic (Figure 5d) implies that the sensitivity of Gsi to D decreased with increasing soil moisture.

Figure 5.

Sensitivity of canopy stomatal conductance calculated from sap flow (Gsi) to vapor pressure deficit (m in Eq. (8) to reference canopy conductance (GsiR) for eight individuals across (a) five solar radiation (Rs) levels and (b) five relative extractable soil water (θE) levels. (c) and (d) show the influence of Rs and graphic on the actual to theoretical (λat) ratio of m/GsiR, respectively. The vertical dashed line in (d) separates isohydric behavior at lowθE and anisohydric behavior at high graphic.

Figure 5.

Sensitivity of canopy stomatal conductance calculated from sap flow (Gsi) to vapor pressure deficit (m in Eq. (8) to reference canopy conductance (GsiR) for eight individuals across (a) five solar radiation (Rs) levels and (b) five relative extractable soil water (θE) levels. (c) and (d) show the influence of Rs and graphic on the actual to theoretical (λat) ratio of m/GsiR, respectively. The vertical dashed line in (d) separates isohydric behavior at lowθE and anisohydric behavior at high graphic.

GsiR of all individuals increased with Rs and graphic according to commonly observed patterns, but there were large differences in the absolute values of GsiR among individuals (Figure 6a and b). Because Gsi was insensitive to D at Rs < 150 W m−2, Eq. (10) cannot be used to extract GsiR; we thus obtained an approximate value selecting Gsi values falling in the range of D of 0.9–1.1 kPa (see Figure 4) encompassing the value of D at which GsiR is determined (D = 1.0 kPa). We show that these values linearly increased with Rs (Figure 6a), noting that, because at such low light levels Gsi is not affected by D, the linear slope of GsiRs up to ∼200 W m−2 was the same at all levels of D, as observed by Lu et al. (2003).

Figure 6.

Relationship between absolute (a, b) and relative (c, d) values of reference canopy stomatal conductance (GsiR) to (a, c) solar radiation (Rs) and (b, d) relative extractable soil water (graphic) of eight individuals. The relative GsiR was the absolute value normalized by the predicted value from the ratio of sapwood area to leaf area (As/Al, Figure 7). For the relationships of GsiR to Rs the analysis is under > 0.29, and for the relationships of GsiR to under Rs > 220 W m−2. The solid lines in (c) and (d) are exponential growths to a maximum based on the averaged parameters of all the individuals. The dashed line in (c) represents the fit for data with < 0.29. Values with D near one (0.9–1.1 kPa) for data with Rs < 150 W m−2 were shown in (a) and (c). The regression function and averaged parameters are shown in Table 3.

Figure 6.

Relationship between absolute (a, b) and relative (c, d) values of reference canopy stomatal conductance (GsiR) to (a, c) solar radiation (Rs) and (b, d) relative extractable soil water (graphic) of eight individuals. The relative GsiR was the absolute value normalized by the predicted value from the ratio of sapwood area to leaf area (As/Al, Figure 7). For the relationships of GsiR to Rs the analysis is under > 0.29, and for the relationships of GsiR to under Rs > 220 W m−2. The solid lines in (c) and (d) are exponential growths to a maximum based on the averaged parameters of all the individuals. The dashed line in (c) represents the fit for data with < 0.29. Values with D near one (0.9–1.1 kPa) for data with Rs < 150 W m−2 were shown in (a) and (c). The regression function and averaged parameters are shown in Table 3.

At optimal light (Rs > 760 W m−2) and soil moisture (graphic; > 0.35), GsiR ranged from 226.7 to 351.0 mmol m−2 s−1, averaging 292.5 mmol m−2 s−1, with a CV of 15%. Much of the variation of GsiR under these conditions was driven by the variation of Kl (P = 0.01; Figure 7a), which, because the variation of water potential was small over time and among individuals (Table 3), was driven primarily by the variation of As/Al (P < 0.001; Figure 7b). GsiR and Kl might be auto-correlated because they are both calculated from the same El, while the relationship between stomatal conductance and hydraulic conductance might reflect the tight correlation between Gsi and gsi (Figure 3). However, normalizing sap flow by the leaf-to-soil potential gradient, thus producing a vine-averaged hydraulic conductivity, resulted in no relationship with As/Al (R2 = 0.13, P = 0.386), demonstrating that inter-stem variation of Kl is driven by As/Al rather than by driving force or differences in tissue-specific hydraulic conductivity. To account for the effect of As/Al on GsiR, each value of the response of GsiR to Rs andθE (Figure 6a and b) was normalized by an estimate obtained from the individual stem's As/Al and the relationship of GsiR = 35.8 × As/Al + 206.2 (R2 = 0.76; P = 0.005; Figure 7c). The CV of the normalized relationships (8%; Figure 6c and d) was about half that of the original.

Figure 7.

Relationship between reference canopy stomatal conductance (GsiR) under saturating solar radiation (Rs > 760 W m−2) and non-limiting soil relative extractable water (graphic > 0.35) to (a) leaf-specific hydraulic conductance (Kl) and (c) the ratio of sapwood area to leaf area (As/Al) for eight individuals; (b) relationship between Kl and As/Al from different dates; for the date of 27 August, the first and second eight vines were separated. There was no significant difference (P = 0.396) between lines representing different dates. Linear regressions are shown with solid lines.

Figure 7.

Relationship between reference canopy stomatal conductance (GsiR) under saturating solar radiation (Rs > 760 W m−2) and non-limiting soil relative extractable water (graphic > 0.35) to (a) leaf-specific hydraulic conductance (Kl) and (c) the ratio of sapwood area to leaf area (As/Al) for eight individuals; (b) relationship between Kl and As/Al from different dates; for the date of 27 August, the first and second eight vines were separated. There was no significant difference (P = 0.396) between lines representing different dates. Linear regressions are shown with solid lines.

Table 3.

Predawn (ψpd, MPa) and midday leaf water potentials (ψmd, MPa) measured on different days.1

Date 7/21 7/28 8/17 8/27 
ψpd −0.45(0.04)a −0.38(0.02)b — −0.34(0.02)b 
ψmd −1.60(0.03)a −1.54(0.05)a −1.71(0.04)a −1.64(0.05)a 
Δψ 1.15(0.04)a 1.16(0.04)a — 1.30(0.05)a 
Date 7/21 7/28 8/17 8/27 
ψpd −0.45(0.04)a −0.38(0.02)b — −0.34(0.02)b 
ψmd −1.60(0.03)a −1.54(0.05)a −1.71(0.04)a −1.64(0.05)a 
Δψ 1.15(0.04)a 1.16(0.04)a — 1.30(0.05)a 

1Values in parentheses represent 1 SE. Different superscripted letters represent significantly different values (P > 0.05) among measuring dates.

As expected from the decrease in CV, the curves representing individuals following normalization were grouped tightly. The normalized GsiR were analyzed for relationships with Rs and graphic based on an exponential rise to a maximum function:  

(9)
formula

The means and standard deviation of the parameters describing both relationships are provided in Table 4.

Table 4.

Parameters of fit function for reference canopy stomatal conductance (Gsi,ref) vs. solar radiation (Rs) and extractable soil water over a depth of 1.0 m (graphic).1

Parameter  G siR –R s  G siR  − graphic 
y 0  37.4 × 10−3 (12.1 × 10−3
a  1.09 (0.04)  1.06 (0.05) 
b  3.7 × 10−3 (0.3 × 10−3 7.27 (0.70) 
CV of b  9.1%  26.9% 
Parameter  G siR –R s  G siR  − graphic 
y 0  37.4 × 10−3 (12.1 × 10−3
a  1.09 (0.04)  1.06 (0.05) 
b  3.7 × 10−3 (0.3 × 10−3 7.27 (0.70) 
CV of b  9.1%  26.9% 

1All the regressions, expressed as Y = y0 + a × (1− exp (−b × X)), were significant at the P = 0.01 level and with R2 > 0.95. Values in parentheses represent 1 SE.

GsiR reached saturation (defined as >90% of the maximum value) in seven of the eight individuals as Rs increased above 400 W m−2 (Figure 6c). GsiR of one individual increased at a slower rate, reaching saturation at an Rs of ∼580 W m−2; this individual displayed the highest L of any row section (2.57 m2 m−2), and might have experienced more self-shading than the others.

Between irrigations, soil moisture decreased fairly quickly (Figure 2), yet GsiR was relatively insensitive until relative extractable water decreased to about a third of the maximum (graphic ∼ 0.35, Figure 6b and d). We note that the effects of Rs and graphic on GsiR were not interactive; for example, lower within the limiting range caused a proportional reduction of GsiR along the entire range of Rs (Figure 6c).

The hypothesis that the rate of decrease in soil moisture will be higher in sections of vineyard rows where Li is high was supported by the results from the oven-dried soil samples extracted near each individual in the last two drying cycles (Figure 8a and b). Normalized by the decrease in vineyard-averaged graphic, the decrease of graphic during each drying cycle—largely representing the rate of soil moisture extraction by individual vines—was 10–20% less than the vineyard average in row sections of low Li, exponentially increasing to 20–30% more than average where Li was higher than average (Figure 8a and b). Depletion of soil moisture rate between irrigation events progressed 36% faster (P < 0.001) in three locations where average Li = 1.6 m2 m−2 than in the other three where average Li = 2.3 m2 m−2. Based on these results, the variation observed in the rate of decline of As/Al-normalized GsiR with graphic represents local variation ofθEi driven by variation of Li. The parameter −b in Eq. (11) represents the sensitivity of GsiR to graphic. We found that −b increased with Li, demonstrating that GsiR decreased faster with graphic as the local Li increased (Figure 8c; piecewise least-squares regression P < 0.001).

Figure 8.

Relationship between decrease of soil volumetric water content (θ) near each vine and leaf area index of a 2-m-long section (1 m to either side of the sap flow vine, Li) during the drying cycles after the fourth (a) and fifth (b) irrigation; (c) relationship between parameter −b in Eq. (9), the sensitivity of reference canopy stomatal conductance (GsiR) to relative extractable soil water over a depth of 1.0 m (graphic), and Li. The dashed line in (a) and (b) is the average decreasing percentage of all the individuals. The solid line in (a) and (b) is the exponential increase fit curve, and in (c) is the linear piecewise fit line.

Figure 8.

Relationship between decrease of soil volumetric water content (θ) near each vine and leaf area index of a 2-m-long section (1 m to either side of the sap flow vine, Li) during the drying cycles after the fourth (a) and fifth (b) irrigation; (c) relationship between parameter −b in Eq. (9), the sensitivity of reference canopy stomatal conductance (GsiR) to relative extractable soil water over a depth of 1.0 m (graphic), and Li. The dashed line in (a) and (b) is the average decreasing percentage of all the individuals. The solid line in (a) and (b) is the exponential increase fit curve, and in (c) is the linear piecewise fit line.

Synthesis of results

Prediction of Gsi was based on the relationships observed with environmental variables (D, Rs and graphic), a hydraulic property (As/Al) and a canopy characteristic (Li). The general formulation was  

(10)
formula
where GsiR was estimated based on As/Al (Figure 7c), λat was estimated from graphic (Figure 5d), the two parameters for f1(Rs) were the average of those obtained from all individuals of the first group (Figure 6c, Table 4), and the two parameters for f2(graphic), specific for each individual, were estimated from Li (Figure 8c).

The pattern of end-of-season departure between modeled and actual values for the first eight-individual group reflects ∼15% non-functional leaf area in early September and ∼35% in late September–early October (Figure 9). One would expect that if the effect of Li on soil moisture depletion was large, individuals growing in sections of high Li would initially show a faster decline of Gsi relative to the average than individuals growing in low Li sections. However, it appears that during this period, lower transpiration rates (driven by lower functional leaf area, D and Rs) rendered the effect of Li on soil moisture depletion, and thus conductance, less important. Thus, for calculating Gs of the second, independent group of eight vines, the end-of-season pattern of functional Li was incorporated into the calculations for predictions of Gsi of the independent group (hereafter referred to as Gsi′).

Figure 9.

Seasonal pattern of the ratio of daily sap-flux-scaled canopy stomatal conductance (GsA) to predicted GsP (with the model not accounting for leaf chlorosis) for the first eight individuals. Error bars are standard deviations of all the individuals. The vertical dashed line is the day in which leaf chlorosis became noticeable in the vineyard.

Figure 9.

Seasonal pattern of the ratio of daily sap-flux-scaled canopy stomatal conductance (GsA) to predicted GsP (with the model not accounting for leaf chlorosis) for the first eight individuals. Error bars are standard deviations of all the individuals. The vertical dashed line is the day in which leaf chlorosis became noticeable in the vineyard.

Hourly Gsi′ values predicted based on this approach were linearly related to measured values (see means and SD in Figure 10a). The residuals were unrelated to D, Rsgraphic, or Li (data not shown) and the root mean square error range among individuals was 22.37–37.28 mmol m−2 s−1, averaging 29.07 mmol m−2 s−1. This is reflected in the closeness of the modeled–measured relationships of all individuals to the unity line (Figure 10b); the intercepts of the individual relationships were significantly higher than but absolutely close to zero (ranging from 1.20 to 14.00 mmol m−2 s−1, averaging 7.10 ± 3.57 mmol m−2 s−1, P for difference from zero = 0.001), and the slopes were close to one (ranging from 0.92 to 1.05, averaging 0.97 ± 0.05, P = 0.113).

Figure 10.

Comparison between modeled canopy stomatal conductance (Gsi′) and sap-flux scaled Gs for the second eight individuals during the period of 13 August to 5 October. (a) shows the averaged hourly data and standard deviation of all the individuals and (b) shows the slopes of modeled Gs to sap-flux scaled Gs of each individual.

Figure 10.

Comparison between modeled canopy stomatal conductance (Gsi′) and sap-flux scaled Gs for the second eight individuals during the period of 13 August to 5 October. (a) shows the averaged hourly data and standard deviation of all the individuals and (b) shows the slopes of modeled Gs to sap-flux scaled Gs of each individual.

Discussion

The motivation for this study was to quantify the contributions of plant and canopy characteristics to spatiotemporal variation of Gsi in vineyards of arid regions. Information leading to optimization of crop water use with respect to yield can provide water resource managers with the tools necessary to sustain agriculture in water-limited environments. We showed at the individual vine scale that temporal responses of GsiR to light and D (Figures 4 and 6a and c) are very similar to those previously observed by Lu et al. (2003), and mostly predictable based on the hydraulic theory that stomatal response to D should be consistent with protecting the hydraulic system from failure (Tyree and Ewers 1991, Oren et al. 1999, Sperry 2000, Sperry et al. 2002, Brodribb 2009). Although these results are confirmatory, together with the finding that the only individual not reaching complete light saturation was positioned in the row section of the highest Li, and the similarity of observations in conductance obtained from sap flow and from scaled leaf-level gas-exchange measurements (Figure 3; see similar results in Ewers et al. 2005, Kim et al. 2008), they lend confidence to the more novel findings of this work.

First, our scant data on leaf water potential (Table 3) suggest a stomatal behavior that maintains an approximately constant water potential gradient driving water from soil to leaf regardless of soil moisture conditions (isohydrodynamic behavior; Franks et al. 2007). However, our analyses of vine-averaged stomatal conductance suggest a transition from a behavior designed to protect the integrity of the hydraulic system through tight stomatal regulation of transpiration (isohydric) under soil water limitation conditions to a less regulating behavior (anisohydric) under soil with high moisture availability. The expected stomatal regulation for isohydric behavior straddles the unity line of the ratio of actual to expected stomatal conductance sensitivity to D (Figure 5d; Oren et al. 1999, Ewers et al. 2001, Ewers et al. 2005, Herbst et al. 2007, 2008, Kim et al. 2008). Such isohydric behavior is associated with a ratio of m to GsiRa) close to the theoretical ratio (λt), seen in Figure 5d when soil water is low, while λa lower than λt indicates an anisohydric behavior, progressively more noticeable in Figure 5d as soil moisture increased. Lacking higher frequency data on leaf water potential and the dynamics of the hydraulic system, we cannot resolve the inconsistency between an apparent isohydrodynamic behavior observed in the water potential data and an apparent anisohydric–isohydric transitional behavior observed in the conductance data.

Certain varieties of V. vitifera are isohydric and others are anisohydric (Schultz and Stoll 2010). Schultz (2003) found a variety of mesic origin (Syrah) showing anisohydric stomatal behavior, whereas another variety (Grenache), which originates from the Mediterranean basin, showed near-isohydric behavior. Furthermore, one variety (Tempranillo) showed isohydric behavior in one study and anisohydric behavior in other studies (Flexas et al. 1998, Medrano et al. 2003), and another variety (Pinot noir) showed divergent behavior in different drought treatments of the same study (Poni et al. 1993, Lovisolo et al. 2010). However, this study on a fifth variety (Merlot) is the first to demonstrate a changed behavior from anisohydric under wet soil to isohydric when soil water depleted over drying cycles. This suggests that tight stomatal regulation (isohydric) is related to lower soil moisture, consistent with the finding that stomatal sensitivity to D in an anisohydric grape variety (Syrah) increased under deficit irrigation (Collins et al. 2010). A study on the Semillon grape variety suggests that anisohydric behavior might be related to, or reflected in, higher night-time sap flow (Rogiers et al. 2009), often correlated with higher soil moisture conditions (Phillips and Oren 1998). The different stomatal behavior observed in studies of certain varieties depending on soil moisture conditions (Poni et al. 1993, Flexas et al. 1998, Medrano et al. 2003, Collins et al. 2010, Lovisolo et al. 2010), together with our results, suggests that the two neatly distinguishable classifications of stomatal regulation may not represent well the actual stomatal behavior. Nevertheless, studies on Arabidopsis and tomato in which genes were infused to alter isohydric to anisohydric behavior resulted in a greatly enhanced growth of plants, but lower drought tolerance (Sade et al. 2009, 2010). If the transitional behavior suggested by our analysis is verified through more direct measurements, it may represent an ideal balance between production when soil moisture is available and survival when soil moisture is more depleted.

Another finding we wish to highlight is the effect of variation in sapwood area per unit of leaf area, an individual vine hydraulic property, on Gsi (Figure 7). The variable As/Al contributes to leaf-specific hydraulic conductance (Kl), a lumped variable shown in many studies to explain much of the variation of stomatal conductance under standard atmospheric and soil conditions (Hubbard et al. 2001, Comstock 2002, Ward et al. 2008). The other contributing variables to Kl are the path length for water transport (assumed here to be similar for all individuals), the water potential gradient driving the flow and an integrated path-length hydraulic specific conductivity. The force driving the flow was similar among individuals (observe the small variance of Δψ in Table 3), and thus normalizing the flow by the driving force did not explain the variation of Gsi and did not change significantly (P = 0.605) from wet to dry soil conditions. Thus, in this case, the variable dominating the relationship between Gsi and Kl (Figure 7a) was As/Al, as demonstrated by the relationship between As/Al and Kl (Figure 7b) and more directly between Gsi and As/Al (Figure 7c). In fact, approximately half of the variation of Gsi among individuals for a particular set of environmental conditions was generated by differences in As/Al (compare Figure 6a–d, respectively). In this fertilized vineyard, spatial variation of sapwood area (CV = 24%), leaf area (30%) and Li (21%) could conceivably be generated by variation in soil physical characteristics. However, saturated soil moisture and parameters of soil moisture–tension relationships were fairly uniform (CV =  ∼ 10%), indicating that the variability of Gs was more likely caused by other drivers of the spatial variability of As/Al. Indeed, the variability of As/Al was generated by summer canopy hedging performed to shape the canopy of the entire row rather than to keep a uniform L or Al. Thus, not only As/Al varied among individuals, but small sections along rows (2 m long in our study) also supported varying L (i.e., Li).

Our final finding is a process that produces spatiotemporal dynamics in canopy stomatal conductance in most ecological systems, but was particularly enhanced by the linear architecture of the vineyard. In most systems, spatial differences in water use, driven by variation in L, canopy conductance and distribution of roots operate to create spatial variation of soil moisture, even in relatively homogeneous soils. However, hydraulic redistribution of moisture through both soils (Katul et al. 1997, Katul and Siqueira 2010) and roots (Smart et al. 2005, Domec et al. 2010) counters this by operating to homogenize soil moisture. This reduces the range of spatial variation of soil moisture, making it difficult to identify feedbacks such as (i) a higher than average water use in certain spots when soil moisture is high (e.g., higher transpiration where L is high), leading to (ii) a faster than average decrease of soil moisture and (iii) a lower than average water use as the drying cycle lengthens (e.g., lower transpiration and conductance where L is high). In our vineyard, most fine root length (82%) was found within 0.8 m from the vine rows on the irrigation furrow side (L. Fang, personal communication). The linear architecture of vineyards should limit the hydraulic redistribution essentially to two dimensions, enhancing spatial variability relative to more typical three-dimensional systems.

Indeed, during two drying cycles with available data, the rate of relative extractable soil moisture depletion near each vine (θEi) was explained to a large extent by Li (Figure 8a and b). Furthermore, the values of the coefficient describing the rate of decrease of Gsi with decreasing graphic were similarly related to Li (Figure 8c). Thus, the data demonstrate that row sections in which L is higher were associated with a faster decrease of available moisture, leading to a faster decrease in Gsi. Although the process is more readily discernible in the vineyard, its effect, e.g., an inverse relationship between a surrogate of Li and sap flux density, has been observed, albeit less clearly, in forests (Oren et al. 1998). Considering that most models of canopy-level stomatal conductance are one dimensional, thus assuming planar homogeneity of soil moisture availability, the non-linearity of the observed response (Figure 7) demonstrates that, in systems where such an assumption is inappropriate, estimates of stomatal conductance may be progressively overestimated as droughts worsen. This was demonstrated when the effect on soil moisture and conductance was modeled with two canopies of the same average L but different L distributions (one with an even L of 1.9 m2 m−2, the other composed of half the vineyard at L = 1.2 m2 m−2 and the other half at 2.6 m2 m−2, the lowest and highest Li observed). The ∼12% reduction in average conductance of the spatially variable canopy, associated with ∼6% less relative extractable soil moisture (Figure 11), would translate to a similar reduction in photosynthesis (Jarvis and Mansfield 1981) during the fruit production period, likely affecting yield.

Figure 11.

Seasonal pattern of the ratio of canopy stomatal conductance (Gs) and soil relative extractable water (graphic) modeled for a canopy of even Li (1.9 m2 m−2) and that of a similar average Li but half representing the lowest and half the highest sectional Li observed in the vineyard (1.2 and 2.6 m2 m−2), respectively.

Figure 11.

Seasonal pattern of the ratio of canopy stomatal conductance (Gs) and soil relative extractable water (graphic) modeled for a canopy of even Li (1.9 m2 m−2) and that of a similar average Li but half representing the lowest and half the highest sectional Li observed in the vineyard (1.2 and 2.6 m2 m−2), respectively.

Regulated deficit irrigation was shown to conserve water in vineyards without adversely affecting yield (Stikic et al. 2003, Kang and Zhang 2004, Fuentes 2006, Chaves et al. 2007). The results of this study suggest that viticultural practices, such as canopy leaf area management, offer a cheap and less technically demanding water-saving alternative.

Funding

This study was supported by the Key Project of the National Natural Science Foundation of China (50939005), the 863 National High Technology Research and Development Program of China (2011AA100502).

Acknowledgments

We would like to thank Jean-Christophe Domec and Danielle Way for their critical comments on this paper.

References

Bao
S.D.
2000
.
Soil and agricultural chemistry analysis
.
China Agricutural Press
,
Beijing
,
495 p
.
Braun
P.
Schmid
J.
1999
.
Sap flow measurements in grapevines (Vitis vinifera L.) 1. Stem morphology and use of the heat balance method
.
Plant Soil
 
215
:
39
45
.
Brodribb
T.J.
2009
.
Xylem hydraulic physiology: the functional backbone of terrestrial plant productivity
.
Plant Sci
 .
177
:
245
251
.
Campbell
G.S.
Norman
J.M.
1998
.
Introduction to environmental biophysics
.
Springer
,
Berlin
,
286 p
.
Čermák
J.
Kučera
J.
Nadezhdina
N.
2004
.
Sap flow measurements with some thermodynamic methods, flow integration within trees and scaling up from sample trees to entire forest stands
.
Trees-Struct. Funct
 .
18
:
529
546
.
Chang
X.
Zhao
W.
Zhang
Z.
Su
Y.
2006
.
Sap flow and tree conductance of shelter-belt in arid region of China
.
Agric. For. Meteorol
 .
138
:
132
141
.
Chaves
M.M.
Santos
T.P.
Souza
C.R.
Ortuno
M.F.
Rodrigues
M.L.
Lopes
C.M.
Maroco
J.P.
Pereira
J.S.
2007
.
Deficit irrigation in grapevine improves water use efficiency while controlling vigour and production quality
.
Ann. Appl. Biol
 .
150
:
237
252
.
Collins
M.J.
Fuentes
S.
Barlow
E.W.R.
2010
.
Partial rootzone drying and deficit irrigation increase stomatal sensitivity to vapour pressure deficit in anisohydric grapevines
.
Funct. Plant Biol
 .
37
:
128
138
.
Comstock
J.P.
2002
.
Hydraulic and chemical signalling in the control of stomatal conductance and transpiration
.
J. Exp. Bot
 .
53
:
195
200
.
Ding
R.
Kang
S.
Li
F.
Zhang
Y.
Tong
L.
Sun
Q.
2010
.
Evaluating eddy covariance method by large-scale weighing lysimeter in a maize field of northwest China
.
Agric. Water Manag
 .
98
:
87
95
.
Domec
J.C.
Gartner
B.L.
2003
.
Relationship between growth rates and xylem hydraulic characteristics in young, mature and old growth ponderosa pine trees
.
Plant, Cell Environ
 .
26
:
471
483
.
Domec
J.C.
Noormets
A.
King
J.S.
Sun
G.E.
Mcnulty
S.G.
Gavazzi
M.J.
Boggs
J.L.
Treasure
E.A.
2009
.
Decoupling the influence of leaf and root hydraulic conductances on stomatal conductance and its sensitivity to vapour pressure deficit as soil dries in a drained loblolly pine plantation
.
Plant, Cell Environ
 .
32
:
980
991
.
Domec
J.C.
Palmroth
S.
Ward
E.
Maier
C.A.
Thérézien
M.
Oren
R.
2009
.
Acclimation of leaf hydraulic conductance and stomatal conductance of Pinus taeda (loblolly pine) to long term growth in elevated CO2 (free air CO2 enrichment) and N fertilization
.
Plant, Cell & Environ
 .
32
:
1500
1512
.
Domec
J.C.
King
J.S.
Noormets
A.
Treasure
E.
Gavazzi
M.J.
Sun
G.
McNulty
S.G.
2010
.
Hydraulic redistribution of soil water by roots affects whole stand evapotranspiration and net ecosystem carbon exchange
.
New Phytol
 .
187
:
171
183
.
Eckardt
N.A.
Cominelli
E.
Galbiati
M.
Tonelli
C.
2009
.
The future of science: food and water for life
.
Plant Cell
 
21
:
368
372
.
Ewers
B.E.
Oren
R.
2000
.
Analyses of assumptions and errors in the calculation of stomatal conductance from sap flux measurements
.
Tree Physiol
 .
20
:
579
589
.
Ewers
B.E.
Oren
R.
Johnsen
K.H.
Landsberg
J.J.
2001
.
Estimating maximum mean canopy stomatal conductance for use in models
.
Can. J. For. Res
 .
31
:
198
207
.
Ewers
B.E.
Gower
S.T.
Bond Lamberty
B.
Wang
C.K.
2005
.
Effects of stand age and tree species on canopy transpiration and average stomatal conductance of boreal forests
.
Plant, Cell Environ
 .
28
:
660
678
.
Ewers
B.E.
Mackay
D.S.
Samanta
S.
2007
.
Interannual consistency in canopy stomatal conductance control of leaf water potential across seven tree species
.
Tree Physiol
 .
27
:
11
24
.
Feng
Q.
Cheng
G.D.
Masao
M.K.
2000
.
Trends of water resource development and utilization in arid north-west China
.
Environ. Geol
 .
39
:
831
838
.
Flexas
J.
Escalona
J.M.
Medrano
H.
1998
.
Down-regulation of photosynthesis by drought under field conditions in grapevine leaves
.
Aust. J. Plant Physiol
 .
25
:
893
900
.
Franks
P.J.
Drake
P.L.
Froend
R.H.
2007
.
Anisohydric but isohydrodynamic: seasonally constant plant water potential gradient explained by a stomatal control mechanism incorporating variable plant hydraulic conductance
.
Plant, Cell Environ
 .
30
:
19
30
.
Fuentes
S.
2006
.
Precision irrigation for grapevines (Vitis vinifera L.) under RDI and PRD
.
Ph.D. thesis, University of Western Sydney
.
284 p
.
Grainer
A.
1987
.
Evaluation of transpiration in a Douglas fir stand by means of sap flow measurement
.
Tree Physiol
 .
3
:
309
320
.
Granier
A.
Loustau
D.
Bréda
N.
2000
.
A generic model of forest canopy conductance dependent on climate, soil water availability and leaf area index
.
Ann. For. Sci
 .
57
:
755
765
.
Grime
V.L.
Morison
J.
Simmonds
L.P.
1995
.
Sap flow measurements from stem heat balances: a comparison of constant with variable power methods
.
Agric. For. Meteorol
 .
74
:
27
40
.
Herbst
M.
Roberts
J.M.
Rosier
P.T.W.
Gowing
D.J.
2007
.
Seasonal and interannual variability of canopy transpiration of a hedgerow in southern England
.
Tree Physiol
 .
27
:
321
333
.
Herbst
M.
Rosier
P.T.W.
Morecroft
M.D.
Gowing
D.J.
2008
.
Comparative measurements of transpiration and canopy conductance in two mixed deciduous woodlands differing in structure and species composition
.
Tree Physiol
 .
28
:
959
970
.
Hubbard
R.M.
Ryan
M.G.
Stiller
V.
Sperry
J.S.
2001
.
Stomatal conductance and photosynthesis vary linearly with plant hydraulic conductance in ponderosa pine
.
Plant, Cell Environ
 .
24
:
113
121
.
Hupet
F.
Vanclooster
M.
2004
.
Sampling strategies to estimate field areal evapotranspiration fluxes with a soil water balance approach
.
J. Hydrol
 .
292
:
262
280
.
Jarvis
P.G.
1976
.
The interpretation of the variations in leaf water potential and stomatal conductance found in canopies in the field
.
Philos. Trans. R. Soc. London B, Biol. Sci
 .
273
:
593
610
.
Jarvis
P.G.
Mansfield
T.A.
1981
.
Stomatal physiology
.
Cambridge University Press
,
Cambridge, UK
,
287 p
.
Jarvis
P.G.
McNaughton
K.G.
1986
.
Stornatal control of transpiration: scaling up from leaf to region
.
Adv. Ecol. Res. 15
 :
1
49
.
Jost
G.
Heuvelink
G.
Papritz
A.
2005
.
Analysing the space-time distribution of soil water storage of a forest ecosystem using spatio-temporal kriging
.
Geoderma
 
128
:
258
273
.
Kang
S.
Zhang
J.
2004
.
Controlled alternate partial root-zone irrigation: its physiological consequences and impact on water use efficiency
.
J. Exp. Bot
 .
55
:
2437
2446
.
Kang
S.
Su
X.
Tong
L.
Shi
P.
Yang
X.
Abe
Y.
Du
T.
Shen
Q.
Zhang
J.
2004
.
The impacts of human activities on the water–land environment of the Shiyang River basin, an arid region in northwest
China. Hydrol. Sci. J
 .
49
:
413
427
.
Kang
S.
Su
X.
Tong
L.
Zhang
J.
Zhang
L.
Davies
W.J.
2008
.
A warning from an ancient oasis: intensive human activities are leading to potential ecological and social catastrophe
.
Int. J. Sustainable Dev. World Ecol
 .
15
:
440
447
.
Katul
G.G.
Siqueira
M.B.
2010
.
Biotic and abiotic factors act in coordination to amplify hydraulic redistribution and lift
.
New Phytol
 .
187
:
3
6
.
Katul
G.
Todd
P.
Pataki
D.
Kabala
Z.J.
Oren
R.
1997
.
Soil water depletion by oak trees and the influence of root water uptake on the moisture content spatial statistics
.
Water Resour. Res
 .
33
:
611
623
.
Kim
H.S.
Oren
R.
Hinckley
T.M.
2008
.
Actual and potential transpiration and carbon assimilation in an irrigated poplar plantation
.
Tree Physiol
 .
28
:
559
577
.
Köstner
B.M.M.
Schulze
E.D.
Kelliher
F.M.
Hollinger
D.Y.
Byers
J.N.
Hunt
J.E.
McSeveny
T.M.
Meserth
R.
Weir
P.L.
1992
.
Transpiration and canopy conductance in a pristine broad-leaved forest of Nothofagus: an analysis of xylem sap flow and eddy correlation measurements
.
Oecologia
 
91
:
350
359
.
Lebon
E.
Dumas
V.
Pieri
P.
Schultz
H.R.
2003
.
Modelling the seasonal dynamics of the soil water balance of vineyards
.
Funct. Plant Biol
 .
30
:
699
710
.
Li
S.
Tong
L.
Li
F.
Zhang
L.
Zhang
B.
Kang
S.
2009
.
Variability in energy partitioning and resistance parameters for a vineyard in northwest
China. Agric. Water Manag
 .
96
:
955
962
.
Liu
M.C.
Zhang
F.
Jiang
J.F.
Wei
Y.G.
2006
.
Influence of climate resources on brewing grape along desert area in Hexi Corridor
.
Agric. Res. Arid Areas
 
24
:
143
148
.
Lovisolo
C.
Perrone
I.
Carra
A.
Ferrandino
A.
Flexas
J.
Medrano
H.
Schubert
A.
2010
.
Drought-induced changes in development and function of grapevine (Vitis spp.) organs and in their hydraulic and non-hydraulic interactions at the whole-plant level: a physiological and molecular update
.
Funct. Plant Biol
 .
37
:
98
116
.
Lu
P.
Yunusa
I.
Walker
R.R.
Muller
W.J.
2003
.
Regulation of canopy conductance and transpiration and their modelling in irrigated grapevines
.
Funct. Plant Biol
 .
30
:
689
698
.
Lu
P.
Urban
L.
Zhao
P.
2004
.
Granier's thermal dissipation probe (TDP) method for measuring sap flow in trees: theory and practice
.
Acta Bot. Sin
 .
46
:
631
646
.
Magnani
F.
Leonardi
S.
Tognetti
R.
Grace
J.
Borghetti
M.
1998
.
Modelling the surface conductance of a broad leaf canopy: effects of partial decoupling from the atmosphere
.
Plant, Cell Environ
 .
21
:
867
879
.
Medrano
H.
Escalona
J.M.
Cifre
J.
Flexas
J.
Bota
J.
2003
.
A ten-year study on the physiology of two Spanish grapevine cultivars under field conditions: effects of water availability from leaf photosynthesis to grape yield and quality
.
Funct. Plant Biol
 .
30
:
607
619
.
Monteith
J.L.
Unsworth
M.H.
1990
.
Principles of environmental physics
,
2nd edition
.
London
:
Edward Arnold
.
Oren
R.
Phillips
N.
Katul
G.
Ewers
B.E.
Pataki
D.E.
1998
.
Scaling xylem sap flux and soil water balance and calculating variance: a method for partitioning water flux in forests
.
Ann. Sci. For
 .
55
:
191
216
.
Oren
R.
Sperry
J.S.
Katul
G.G.
Pataki
D.E.
Ewers
B.E.
Phillips
N.
Schäfer
K.
1999
.
Survey and synthesis of intra- and interspecific variation in stomatal sensitivity to vapour pressure deficit
.
Plant, Cell Environ
 .
22
:
1515
1526
.
Oren
R.
Sperry
J.S.
Ewers
B.E.
Pataki
D.E.
Phillips
N.
Megonigal
J.P.
2001
.
Sensitivity of mean canopy stomatal conductance to vapor pressure deficit in a flooded Taxodium distichum L. forest: hydraulic and non-hydraulic effects
.
Oecologia
 
126
:
21
29
.
Phillips
N.
Oren
R.
1998
.
A comparison of two daily representations of canopy conductance based on two conditional time-averaging methods and the dependence of daily conductance on environmental factors
.
Ann. Sci. For
 .
55
:
217
235
.
Poni
S.
Lakso
A.N.
Turner
J.R.
Melious
R.E.
1993
.
The effects of pre- and post-veraison water stress on growth and physiology of potted Pinot Noir grapevines at varying crop levels
.
Vitis
 
32
:
207
214
.
Rogiers
S.Y.
Greer
D.H.
Hutton
R.J.
Landsberg
J.J.
2009
.
Does night-time transpiration contribute to anisohydric behavior in a Vitis vinifera cultivar?
J. Exp. Bot
 .
60
:
3751
3763
.
Sade
N.
Vinocur
B.J.
Diber
A.
Shatil
A.
Ronen
G.
Nissan
H.
Wallach
R.
Karchi
H.
Moshelion
M.
2009
.
Improving plant stress tolerance and yield production: is the tonoplast aquaporin SlTIP2;2 a key to isohydric to anisohydric conversion?
New Phytol
 .
181
:
651
661
.
Sade
N.
Gebretsadik
M.
Seligmann
R.
Schwartz
A.
Wallach
R.
Moshelion
M.
2010
.
The role of tobacco aquaporin1 in improving water use efficiency, hydraulic conductivity, and yield production under salt stress
.
Plant Physiol
 .
152
:
245
254
.
Sakuratani
T.
1981
.
A heat balance method for measuring water flux in the stem of intact plants
.
J. Agric. Meteorol
 .
37
:
9
17
.
Schäfer
K.
Oren
R.
Tenhunen
J.D.
2000
.
The effect of tree height on crown level stomatal conductance
.
Plant, Cell Environ
 .
23
:
365
375
.
Schultz
H.R.
2003
.
Differences in hydraulic architecture account for near-isohydric and anisohydric behaviour of two field-grown Vitis vinifera L. cultivars during drought
.
Plant, Cell Environ
 .
26
:
1393
1405
.
Schultz
H.R.
Stoll
M.
2010
.
Some critical issues in environmental physiology of grapevines: future challenges and current limitations
.
Aust. J. Grape Wine Res
 .
16
:
4
24
.
Smart
D.R.
Carlisle
E.
Goebel
M.
Núñez
B.A.
2005
.
Transverse hydraulic redistribution by a grapevine
.
Plant, Cell Environ
 .
28
:
157
166
.
Soar
C.J.
Speirs
J.
Maffei
S.M.
Penrose
A.B.
McCarthy
M.G.
Loveys
B.R.
2006
.
Grape vine varieties Shiraz and Grenache differ in their stomatal response to VPD: apparent links with ABA physiology and gene expression in leaf tissue
.
Aust. J. Grape Wine Res
 .
12
:
2
12
.
Sperry
J.S.
2000
.
Hydraulic constraints on plant gas exchange
.
Agric. For. Meteorol
 .
104
:
13
23
.
Sperry
J.S.
Hacke
U.G.
Oren
R.
Comstock
J.P.
2002
.
Water deficits and hydraulic limits to leaf water supply
.
Plant, Cell Environ
 .
25
:
251
263
.
Steduto
P.
Hsiao
T.C.
1998
.
Maize canopies under two soil water regimes. II. Seasonal trends of evapotranspiration, carbon dioxide assimilation and canopy conductance, and as related to leaf area index
.
Agric. For. Meteorol. 89
 :
185
200
.
Stikic
R.
Popovic
S.
Srdic
M.
Savic
D.
Jovanovic
Z.
Prokic
L.J.
Zdravkovic
J.
2003
.
Partial root drying (PRD): a new technique for growing plants that saves water and improves the quality of fruit
.
Bulg. J. Plant Physiol
 .,
Special Issue
.
164
171
.
Stöhr
A.
Lösch
R.
2004
.
Xylem sap flow and drought stress of Fraxinus excelsior saplings
.
Tree Physiol
 .
24
:
169
180
.
Trambouze
W.
Voltz
M.
2001
.
Measurement and modelling of the transpiration of a Mediterranean vineyard
.
Agric. For. Meteor
 .
107
:
153
166
.
Tyree
M.T.
Ewers
F.W.
1991
.
The hydraulic architecture of trees and other woody plants
.
New Phytol
 .
119
:
345
360
.
Uddling
J.
Pleijel
H.
Karlsson
P.E.
2004
.
Measuring and modelling leaf diffusive conductance in juvenile silver birch
,
Betula pendula
 .
Trees-Struct. Funct
 .
18
:
686
695
.
Vrörösmarty
C.J.
Green
P.
Salisbury
J.
Lammers
R.B.
2000
.
Global water resources: vulnerability from climate change and population growth
.
Science
 
289
:
284
288
.
Ward
E.J.
Oren
R.
Sigurdsson
B.D.
Jarvis
P.G.
Linder
S.
2008
.
Fertilization effects on mean stomatal conductance are mediated through changes in the hydraulic attributes of mature Norway spruce trees
.
Tree Physiol
 .
28
:
579
596
.
Whitehead
D.
1998
.
Regulation of stomatal conductance and transpiration in forest canopies
.
Tree Physiol
 .
18
:
633
644
.
Williams
L.E.
Baeza
P.
2007
.
Relationships among ambient temperature and vapor pressure deficit and leaf and stem water potentials of fully irrigated, field-grown grapevines
.
Am. J. Enol. Vitic
 .
58
:
173
.
Williams
L.E.
Phene
C.J.
Grimes
D.W.
Trout
T.J.
2003
.
Water use of mature Thompson Seedless grapevines in California
.
Irrig. Sci
 .
22
:
11
18
.
Winkel
T.
Rambal
S.
1990
.
Stomatal conductance of some grapevines growing in the field under a Mediterranean environment
.
Agric. For. Meteorol
 .
51
:
107
121
.
Wullschleger
S.D.
Wilson
K.B.
Hanson
P.J.
2000
.
Environmental control of whole-plant transpiration, canopy conductance and estimates of the decoupling coefficient for large red maple trees
.
Agric. For. Meteorol
 .
104
:
157
168
.
Yunusa
I.
Walker
R.R.
Lu
P.
2004
.
Evapotranspiration components from energy balance, sapflow and microlysimetry techniques for an irrigated vineyard in inland Australia
.
Agric. For. Meteorol
 .
127
:
93
107
.
Zhang
B.Z.
Kang
S.Z.
Zhang
L.
Du
T.S.
Li
S.E.
Yang
X.Y.
2007
.
Estimation of seasonal crop water consumption in a vineyard using Bowen ratio-energy balance method
.
Hydrol. Processes
 
21
:
3635
3641
.
Zhang
B.Z.
Kang
S.Z.
Li
F.S.
Zhang
L.
2008
.
Comparison of three evapotranspiration models to Bowen ratio-energy balance method for a vineyard in an arid desert region of northwest China
.
Agric. For. Meteorol
 .
148
:
1629
1640
.
Zhang
Y.
Kang
S.
Ward
E.J.
Ding
R.
Zhang
X.
Zheng
R.
2011
.
Evapotranspiration components determined by sap flow and microlysimetry techniques of a vineyard in northwest China: dynamics and influential factors
.
Agric. Water Manag
 .
98
:
1207
1214
.