Water stress causes stomatal closure but does not reduce canopy evapotranspiration in almond

Agricultural Water Management 168 (2016) 11–22

Contents lists available at ScienceDirect

Agricultural Water Management journal homepage: www.elsevier.com/locate/agwat

Water stress causes stomatal closure but does not reduce canopy evapotranspiration in almond Gerardo M. Spinelli a , Richard L. Snyder b , Blake L. Sanden c , Ken A. Shackel a,∗ a b c

University of California, Davis, Dept. of Plant Sciences, One shields Ave., Davis, CA 95616-8683, USA University of California, Davis, Dept. of Land, Air and Water Resources, One shields Ave., Davis, CA 95616-8683, USA University of California Cooperative Extension, 1031 S. Mt. Vernon Ave., Bakersfield, CA 93307, USA

a r t i c l e

i n f o

Article history: Received 23 April 2015 Received in revised form 3 January 2016 Accepted 8 January 2016 Keywords: Stomatal conductance Photosynthesis Stem water potential Crop coefficient Canopy conductance Eddy covariance

a b s t r a c t We investigated the degree of physiological control of transpiration at the leaf and canopy-level in commercial almond orchards in California during periods of water stress ranging from −0.8 to −2.0 MPa of stem water potential. Field measurement of evapotranspiration using the residual of the energy balance method failed to detect a decrease of actual evapotranspiration (ETa ) relative to standardized reference evapotranspiration for short canopies (ETo ) during weekly periods of water stress. Although a decline in the ratio Ka = ETa /ETo is expected in presence of water stress as a result of stomatal regulation, a flat response to water stress was observed for both daily and average midday Ka . A substantial decrease in Ka was observed after harvest, perhaps caused by the decrease in leaf area resulting from harvest operations. Weekly cycles of irrigation caused a clear reduction in midday stem water potential, a sensitive indicator of water status in plants. At the leaf level, there was a continuous decline in stomatal conductance with declining stem water potential for both sunlit and shaded leaves. Shaded leaves had lower conductance values and a flatter slope than sunlit leaves. The results suggest that a decrease in conductance of roughly 50% is expected at the leaf level for the observed range of stem water potential. A scaling-up exercise together with a model of the energy balance indicated the expected magnitude the reduction in Ka for the observed range of stem water potential but did not account for the apparent disconnect between the leaf scale and the canopy scale. Our results underscored the complications associated with estimating plant water stress by measuring field evapotranspiration, especially when grass ETo is used as the reference. From a water management point of view, this study suggests that inducing mild to moderate water stress in almond may not produce substantial water savings. Published by Elsevier B.V.

1. Introduction Almonds are one of the most important agricultural crops in California and a major player in the water budget of the state. Irrigation water is a limiting factor in the Southern San Joaquin Valley, where some farmers received only 20% of their normal water allocation in 2013. In tree crops, mild water stress can be intentionally induced to control excessive vegetative growth (Chalmers et al., 1981) and regulated deficit irrigation can be applied to conserve water (Romero et al., 2004; Girona et al., 2005; Goldhamer et al., 2006). Additionally, in almond production, irrigation is commonly cut off to facilitate harvest and reduce incidence of fungal diseases (Teviotdale et al., 2001). Given widespread water stress due

∗ Corresponding author. E-mail addresses: [email protected] (G.M. Spinelli), [email protected] (R.L. Snyder), [email protected] (B.L. Sanden), [email protected] (K.A. Shackel). http://dx.doi.org/10.1016/j.agwat.2016.01.005 0378-3774/Published by Elsevier B.V.

to both drought and management practice, however, it is important to understand the physiological response and water use of almond orchards under different degrees of imposed water stress. This knowledge will help guide irrigation decisions and increase water use efficiency. To perform photosynthesis, plants face an inexorable tradeoff between carbon assimilation and water loss. The two processes are inextricably linked because they share the same transport pathway, both at the leaf and the canopy scale. The first step of the pathway is quantified by stomatal conductance (gs ) through the stomatal aperture to the surface of the leaf and the second step by aerodynamic conductance (ga ) from the leaf surface to the atmosphere surrounding the leaf. Stomatal conductance is under physiological control, while aerodynamic conductance is a function of wind speed and leaf size. The same principles for leaf transpiration can be applied to a canopy in the commonly used “Big Leaf” approach. In the Big Leaf model, the first step of the pathway is represented by canopy conductance (gc ), a composite function of stomatal conductance,

12

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

leaf area, and soil conductance (Kyaw Tha Paw and Meyers, 1989). Aerodynamic conductance is then modeled as a function of wind speed and canopy characteristics, i.e., height and roughness. (Allen et al., 1998). Plants physiologically respond to mild water stress with stomatal closure, which reduces stomatal conductance over a short response time (Hsiao, 1973). Stomatal closure resulting from water stress has been shown in almonds (Torrecillas et al., 1996). However stomatal conductance also responds to environmental factors. It is affected by light intensity, possibly modified by the severity of water stress (Jones, 1992, p. 148). In general, decreased stomatal conductance should correspond to decreased canopy conductance, even though each leaf in a canopy is differently oriented and hence experiences different light intensity. Thus, the quantitative aggregate response to water stress of total canopy conductance is a function of the relative contributions from a leaf population experiencing a range of light intensities in a rather complex arrangement. Severe stress advances and accelerates leaf senescence (Bradford and Hsiao, 1982), which reduces leaf area and thus canopy conductance over a significantly longer response time. Therefore, physiological response to water stress should decrease canopy conductance and hence canopy water loss. However, some studies reported that canopy conductance does not depend entirely on physiology, vegetation and soil, but also on non-physiological variables (Baldocchi et al., 1991; Todorovic, 1999), although the influence of some of these variables may depend on the level of coupling (Jarvis and McNaughton, 1986; Moreno et al., 1996). The relative importance of ga and gc determines the degree to which a canopy is “coupled” to the atmosphere and how much a proportional decrease in canopy conductance will decrease plant transpiration (Jarvis, 1984). A “decoupled” canopy can minimally control its transpiration rate via stomatal regulation, because the stomatal resistance is relatively unimportant compared to the larger aerodynamic resistance. The role of aerodynamic conductance in determining the sensitivity of transpiration to stomata is well known (Bange, 1953). Jarvis (in Grace et al., 1981) showed the same results for canopies, suggesting that tall canopies such as forests show high canopy conductance and a coupled canopy. In contrast, short canopies like field crops have low aerodynamic conductance and are decoupled. Such a mechanism is less important for carbon assimilation rate because the canopy has a smaller influence on air CO2 than it has on air vapor pressure (Steduto and Hsiao, 1998b). Thus, in a decoupled canopy, stomatal closure may restrict photosynthesis more than it restricts transpiration. The relative importance of carbon assimilation and water loss has been termed “water use efficiency” (WUE), and it is assessed using measures such as carbon assimilation/transpiration (Tanner and Sinclair, 1983), photosynthetic rate per stomatal conductance (Lambers et al., 2008, p. 53), and yield per irrigation water applied in deficit irrigation practices (Goldhamer et al., 2006; Chalmers et al., 1986). It is generally accepted that water stress, by inducing stomatal closure, increases water use efficiency, however Rouhi et al. (2007), in different almond species, showed contrasting patterns depending on the species. The potential for water conservation is greater in tree crops, since they are more coupled to the atmosphere than short, smooth canopy field crops (Fereres and Soriano, 2007). The degree of coupling helps determine the water use efficiency when trees are under water deficit. Irrigation scheduling is commonly managed using a soil water balance approach, where inputs are rain and irrigation and the most important output is crop evapotranspiration (ETc ). The most common method is to estimate ETc using a reference ET (ETo ) and calculate ETc as:

ETc = ETo Kc

(1)

where Kc is an experimentally determined “crop coefficient” (Allen et al., 1998). It is assumed that the ETo represents a large field of 12 cm tall vegetation that is only limited by weather conditions. A modified Penman–Monteith (PM) equation based on generally accepted environmental physics theory and weather station measurements is used to determine ETo (Allen et al., 1998, 2006). The ETo × Kc approach assumes a simple proportionality of ETc to ETo ; however, this approach may be too simplistic because Kc is highly dependent on the local climate and agronomic practices (Allen et al., 1998). Annandale et al. (1994) criticized the ETo × Kc approach, that is based on the assumption of a simple linear relationship between ETc and ETo . In their study, they show that environmental factors affect grass ET differently than crop ET, particularly when the crop is tall and when canopy conductance is low due to ample irrigation. The ETo × Kc approach is commonly used due to widespread availability of Kc values for various crops in the literature (Doorenbos and Pruitt, 1977; Allen et al., 1998). In theory, Kc should be relatively stable in the short term, following a developmental pattern which is mainly affected by canopy growth and crop phenological stage and does not account for crop water stress. In the following equation, ETa is the actual evapotranspiration of a crop, as opposed to ETc that represents the potential unstressed crop ET: ETa = ETo Ka

(2)

where Ka = ETa /ETo is the actual crop coefficient and describes the combined contribution of crop characteristics and crop stress. An eddy covariance system (Burba, 2013) is often used to measure LE, but this method requires specialized personnel and expensive instruments that mostly limit it to research applications. This method has proven consistent with other measures of ETa such as lysimeters (Castellví and Snyder, 2010; Alfieri et al., 2012). Thus, decreased canopy conductance due to water stress should reduce ETa relative to ETo and Ka relative to Kc = ETc /ETo , where ETc is the crop evapotranspiration under standard conditions. An alternative approach is to time irrigation based on plant water status as indicated by measuring midday stem water potential (stem , Shackel, 2011), where stem is closely associated with physiological and production parameters such as stomatal closure, photosynthesis, yield, and fruit quality (McCutchan and Shackel, 1992; Shackel et al., 1997; Naor et al., 2000). According to a recent Almond Board of California survey,  stem is currently being used as an irrigation management tool by 40% of almond growers. This study aims to test the following hypothesis: If a measurable reduction in stomatal conductance results from moderate to severe water stress that develops under standard commercial practices, and if such a reduction aggregates to a measurable decline in canopy ETa , then ETa should decrease relative to ETo and the ratio Ka = ETa /ETo should decrease during water stress. 2. Methods 2.1. Experimental site The experiment was conducted during 2008 through 2011 on a high-yielding (Schellenberg et al., 2012) almond (Prunus dulcis D.A. Webb) orchard at the Paramount Farming Company in the southern San Joaquin Valley near Lost Hills (N 35◦ 30 37 W 119◦ 40 3 ). The almond cultivars were ‘Nonpareil’ (50%) inter-planted with ‘Monterey’ (50%), grafted on Nemaguard rootstocks (Prunus persica). Trees were planted in 1999 with 6.4 × 7.9 m spacing (2 1 3 trees ha−1 ), with a north-south row orientation on a Milham sandy loam (Fine-loamy, mixed, superactive, thermic, Typic Haplargids). Micro-irrigation was applied with two static fan-jet 4 L h−1 microsprinklers per tree that were located north and south of each

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

13

Table 1 CIMIS evapotranspiration, Eddy Covariance evapotranspiration, precipitation and irrigation applied in the four years of the study. Year

CIMIS ETo (mm)

Eddy Co ETa (mm)

Neutron probe ETa (mm)

Irrigation + rain (mm)

DifferenceApplied—Eddy Co ETa (mm)

2008 2009 2010 2011

1274 1488 1377 1322

1341.1 1562.8 1393.7 1278.2

1421 1450 1393 1315

1413 1497 1418 1337

71.9 −65.8 24.3 58.8

Note: in year 2008 data are reported for the period 3/25/08 to 11/11/08 only.

tree. Fertilizers were injected and applied with this system to the area within each tree row so that the remaining space between rows was maintained without irrigation. Mean annual precipitation at the site was 146 mm, mainly during winter during the experiments. The average annual reference evapotranspiration for the location is 1430 mm. From 2008 to 2010, daily mean air temperature ranged from 2.3 to 37.1 ◦ C and daily average soil temperature was 7.7–28.7 ◦ C. Mean midday wind speed was 1.1–5.1 m s−1 with a mean of 2.17 m s−1 . Meteorological data were collected from CIMIS station #146 of the California Irrigation Management and Information System (Snyder and Pruitt, 1992) located 3 km from the study site. Irrigation was scheduled weekly based on soil moisture (measured with neutron probes) and tree water status (measured as stem water potential) for 24 h or 48 h sets. Total precipitation plus irrigation water applied was 1413 mm in 2008, 1497 mm in 2009, 1418 mm in 2010, and 1337 mm in 2011 (Table 1). Irrigation was cut off ∼10 days before harvest to facilitate tree shaking and reduce the incidence of fungal disease. Shaking dates were 18 August 2008, 21 August 2009, 27 August 2010, and 27 August 2011. 2.2. Eddy covariance measurements Eddy covariance is based on measuring of the components of the energy balance equation: Rn = H + G + LE

fan speed was set to fast. To avoid pitfalls discussed by McDermitt (1990), leaf transpiration was not used and stomatal conductance measurements were limited to <60 s to prevent adjustment to the different cuvette conditions. Measurements were taken at midday, concurrent with stem water potentials. The mean of 20 leaves per tree were measured: 10 upper-canopy sunlit leaves and 10 lower-canopy shaded leaves. The internal PAR of the cuvette was set to track the irradiance measured by the external PAR sensor. Occasionally, a delay in tracking of the internal LED and external PAR was observed, leading to measurements being taken at low internal light for sunlit leaves or high light for shaded leaves. To avoid these discrepancies, data collected at internal PPFD higher than 100 ␮mol m−2 s−1 in shaded leaves and at lower than 1600 ␮mol m−2 s−1 for sunlit leaves was discarded. Data were collected on the variety ‘Nonpareil’ at six locations in California. In 2011, the sites were at Lost Hills, (Kern County, the main experimental site), Arbuckle (Colusa County), Madera (Madera County), and Chico (Butte County). In 2012, the sites examined were at Lost Hills, Davis (Yolo County), Madera, Arbuckle and Corning (Tehama County). Trees were chosen at each location to obtain the largest range of stem water potential possible. At each location and date, the gas exchange and stem water potential of five to ten trees were measured at midday. The mean of sunlit and shaded leaves is reported for each tree and date.

(3)

where Rn is net radiation, H is sensible heat flux and G is ground heat flux. Under the assumptions that the canopy has no heat storage capacity and the energy captured or released in metabolic processes is negligible, all energy inputs into the canopy must be balanced by energy outputs. Therefore, LE can be calculated as: LE = Rn − H − G. Then, ETa is determined by dividing LE by the latent heat of vaporization L = 2.45MJkg−1 . Net radiation was measured with a net radiometer (REBS, Q7.1), ground heat flux with heat flux plates (REBS, Inc., HFT3.1), and soil temperature by averaging 4-probe thermocouples (Campbell Scientific, Inc., Tcav). Sensible heat flux was measured using eddy covariance and data from a sonic anemometer (RM Young, Inc., 81000RE) and using the surface renewal method (Shapland et al., 2013) from high frequency temperature measured with a 76.2 micrometer diameter thermocouple (Campbell Scientific, Inc., FW3). Daily actual evapotranspiration (ETa ) was obtained by dividing latent heat flux by the latent heat of vaporization, L. The observed Ka was calculated by dividing ETa by reference evapotranspiration from the nearby CIMIS station. Midday Ka was calculated similarly using the cumulative ET data between 12:00 and 15:00. 2.3. Gas exchange measurements Photosynthetic rate and stomatal conductance were measured with a photosynthesis unit (LICOR 6400) with the following cuvette settings: PPFD was set to track external PAR, with care taken not to alter leaf orientation during the measurement; flow rate was 700 ␮mol s−1 and sample CO2 , 400 ␮mol mol−1 . Humidity inside the cuvette was allowed to track the ambient conditions; leaf temperature was measured with the leaf thermocouple and the cuvette

2.4. Stem water potential and baseline Plant water status was measured as midday stem water potential (stem ) with a pressure chamber (Soilmoisture Equipment Corporation and PMS Instrument Pump-up) weekly or biweekly. One or two lower-canopy, fully expanded, shaded leaves were measured after enclosing the leaves in bags made of plastic film and aluminum foil for at least one hour before measuring. Each sampling date, 30 trees (out of 384) in the plot were measured in 2008 and 2009; 35 trees in 2010; and 40 trees in 2011. An unstressed baseline was calculated from the CIMIS station readings for air vapor pressure deficit, using the method proposed by McCutchan and Shackel (1992) for prune.

2.5. Micrometeorology The energy balance was modeled to obtain the expected canopy ET given the measured leaf stomatal conductance. Under the assumption of a constant latent heat of vaporization (L = 2.45 MJ kg−1 ), ET is expressed as latent heat (LE) flux in units of energy flux density (W m−2 ). A numerical iterative method was used to estimate the surface temperature that solves the energy balance for a particular set of conditions. Since the stomatal conductance data were collected at midday, the energy balance was modeled for the environmental conditions recorded at midday (average between 12:00 and 15:00). Each term of the energy balance equation (Eq. (3)) was expanded to show dependencies on surface temperature and on canopy conductance. Thus, the energy balance equation was written as

14

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

Table 2 Linear regression analysis of the relationship between actual crop coefficient versus stem water potential. Year

Midseason dates (n)

Range in crop coefficient (used in regression)

Range in average stem water potential (MPa) (used in regression)

Slope

SlopeP-value

2008 2009 2010 2011

7 15 17 16

0.88–1.18 0.92–1.22 0.90–1.24 0.97–1.23

−1.7 to −0.8 −1.9 to −0.6 −2.0 to −0.5 −1.4 to −0.6

−0.0055 0.0038 −0.0106 −0.0010

0.6761 0.5931 0.0234 0.8957

(adapted from Jones, 1992; Eqs. (5.2), (5.17), (5.20) and (5.22) and Campbell and Norman, 1998; Eq. (14.1)): Ri − εTs4

=

 1

Cp esat(Tl) − ea 

gc

+

 1



+ Cp ga (Ts − Ta ) + G

(5)

ga

where Ri is input radiation, the sum of incoming shortwave and longwave radiation minus the reflected shortwave radiation and εsTs 4 represents the longwave radiation emitted by the surface as a function of surface temperature. Since surface temperature changes when gc changes, Ri and not Rn should stay fixed as gc changes (Paw and Gao, 1988). Thus, we recalculated Ri from the measured Rn and used it to model the energy balance.  is air density, Cp is air heat capacity at constant pressure and  is the psychrometric constant. The saturated vapor pressure at leaf temperature (esat(Tl) ) was calculated using both Buck’s (1981) equation and Kyaw (1987), obtaining similar results. Air vapor pressure (ea ) was obtained from the nearby CIMIS station and air temperature (Ta ) in the orchard was obtained from the thermocouples above the canopy that were used in the surface renewal method. In the model, ground heat flux density (G) was estimated as 10% of Rn . The aerodynamic conductance (ga ) was calculated from wind speed (uz ) and friction velocity (u* ) using an equation from Monteith and Unsworth (1990): ga =

u2∗ uz

(6)

Stomatal conductance measurements were scaled up to obtain the expected whole-canopy conductance based on plant water status using the average stem water potential of the orchard for each date. From the mean stem water potential on each date, stomatal conductance of sunlit leaves (gsun ) and of shaded leaves (gsh ) was estimated using a smoothed spline function with 60% smoothing (proc transreg, SAS 9.3, SAS Institute Inc., Cary, NC, USA) from the relationship between stomatal conductance and stem water potential obtained from the gas exchange measurements. Canopy bulk conductance (gc ) was calculated with two methods. The first assumed that half of the leaf area index (LAI) was active and exhibited the stomatal conductance of sunlit leaves using the equation proposed by Allen et al. (1998): gc = gsun 0.5 LAI

(7)

The second method used the mean of sunlight and shaded leaf conductance weighted by the respective LAI (Sinclair et al., 1976) to obtain canopy conductance (gc ). gc = LAIsu gsun + LAIsh gsh

(8)

A constant LAI was assumed for each tree and season, thus, applying a reference value of 3.25 (Zarate-Valdez et al., 2012).Two scenarios were modeled. First, an unstressed ET obtained using the expected gs for a well-watered tree (estimated from the stem ) and scaling up from gs to gc . The canopy conductance obtained is gc unstr . The second scenario modeled the expected ET based on the observed mean water stress of the orchard. The canopy conductance obtained is gc  . To estimate the proportional reduction in LE of the stressed to the unstressed orchard on any given day, we

calculated the ratio between LE based on stem and LE unstressed. This ratio reduces to: / + 1 + ga /gc unstr LESWP = LEunstr / + 1 + ga /gc 

(9)

The observed Ka was multiplied by this ratio to estimate the expected reduction in Ka during days of water stress. 3. Results 3.1. Evapotranspiration and Ka The ETa followed the same general pattern as ETo during 2010 (Fig. 1). Initially, ETa started lower than ETo , but it increased sharply in March when the trees leafed out and exceeded ETo , reaching∼8 mm day−1 during the summer months. The Ka was as low as 0.60 before leaf-out and increased sharply to 1.00 in March and April (Fig. 2). The Ka increased gradually during summer months to 1.20. Both ETa and ETo dropped at the end of August 2010, but ETa dropped more substantially to Ka = 0.75. A similar pattern was observed in 2008, 2009 and 2011, although the late season Ka was somewhat different each year (Fig. 2). 3.2. Irrigation and stem water potential Irrigation, applied approximately once a week, caused cycles of stem water potential in trees of −0.5 to −2.0 MPa depending on irrigation events. All trees within the orchard showed a similar pattern, but at any given time there was a range of stem of ∼0.5 MPa on wet days to 1 MPa on dry days (see SE’s in Fig. 3a). However, there was no reduction in Ka associated with water stress. For instance, on the three driest days in 2010 (29 June, 10 August, and 24 August) the Ka was the same as when the trees were close to baseline (6 July and 27 July). During the midseason, between 1 May and 27 August, the Ka was relatively stable during all four years (Fig. 2, Table 2). During mid season of 2008, 2009, and 2011, regressions between the daily Ka and average stem of all trees on that day yielded slope statistically indistinguishable from zero. In 2010, the slope was significant but negative, indicating an increase in Ka associated with more negative stem , however the magnitude of the slope was −0.01. The same lack of response of Ka to stem was observed for the midday Ka (average of the period 12:00–15:00, Fig. 3b). 3.3. Gas exchange measurements Stomatal conductance showed a robust correlation with stem across sites and years (Fig. 4a, b) with a highly significant slope different from zero and an adjusted R2 = 0.74 for sunlit and R2 = 0.43 for shaded leaves (Table 3). Stomatal conductance for sunlit and shaded leaves ranged from 22 to 530 mmol m−2 s−1 and from 7 to 354 mmol m−2 s−1 , respectively (Fig. 4 a, b). Photosynthetic rate showed a highly significant linear relation to stem for sunlit leaves (Fig. 4c), with a linear adjusted R2 = 0.61 (Table 3), but there was no relation between photosynthesis and stem for shaded leaves (Fig. 4c,Table 3). Photosynthetic rate ranged from 3.58 to 28.05 and from −3.56 to 16.00 for sunlit and shaded leaves, respectively (Fig. 4 c,d).

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

15

Fig. 1. Daily mean values of ETa (black dot), ETo (dashed line) and Ka (solid line) in 2010.

Fig. 2. Daily mean values of Ka in 2008, 2009, 2010 and 2011.

Table 3 Linear regression analysis of the relationship between gas exchange measurements and stem water potential. Gas exchange measurement

Leaf category

Adjusted R2

n

Slope

Slope P-value

Stomatal conductance

Sunlit Shaded

0.74 0.43

172 161

0.017 0.012

<0.0001 <0.0001

Photosynthetic rate

Sunlit Shaded

0.61 −0.0062

172 161

0.802 0.004

<0.0001 0.917

16

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

Fig. 3. (a) Seasonal pattern of midday stem water potential (black line with error bars) and SWP baseline (black circles) in MPa, daily average Ka (black dots) and irrigation applied in decimeters (black squares). The vertical dashed line represents the date of harvest. N was 35 for each day, except for Sep28, 2010, when N was 27. Error bars are +/−2 SE. (b) Seasonal pattern of midday stem water potential (black line with error bars) and SWP baseline (black circles) in MPa, midday average Ka (black dots) and irrigation applied in decimeters (black squares). The vertical dashed line represents the date of harvest. N was 35 for each day, except for Sep28, 2010, when N was 27. Error bars are +/−2 SE.

3.4. Statistical analysis Linear correlations of stomatal conductance and photosynthetic rate to stem were evaluated using proc reg, of the SAS statistical software (SAS 9.3, SAS Institute Inc., Cary, NC, USA). These analyses were carried out with the intent of exploring whether or not the relationship between the variables were statistically significant. For the modeling exercise (see next section), to obtain the correspondent stomatal conductance expected for any value of stem , we used a smooth spline obtained with proc transreg (SAS 9.3, SAS Institute Inc., Cary, NC, USA). The splines obtained show a non-

linear behavior and are shown in Fig. 4. The shape of the splines is sigmoidal, showing flatter response of both stomatal conductance and photosynthetic rate at low and high values of stem , and a steeper response in the range of stem between −1.0 and −1.5 MPa. 3.5. Energy balance model Scaling-up stomatal conductance for the unstressed tree yielded a gc between 21.7 mm s−1 and 23.3 mm s−1 (for rc = 43 and 46 s m−1 ). These values are consistent with the rc = 50 s m−1 used for the daytime hourly calculation of ETo in FAO56 (Allen et al.,

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

17

Fig. 4. (a, b) Relationship between leaf-level stomatal conductance and stem water potential for sunlit (panel a) and shaded leaves (panel b), measured at midday across two years and six orchards. Each symbol represents the average of 5–10 leaves in a tree, the line was obtained with a smoothed spline function with 60% smoothing (proc transreg, SAS 9.3). (c, d) Relationship between leaf-level photosynthetic rate and stem water potential for sunlit (panel c) and shaded leaves (panel d) measured at midday across two years and six orchards. Each symbol represents the average of 5–10 leaves in a tree, the line was obtained with a smoothed spline function with 60% smoothing (proc transreg, SAS 9.3).

1998). The scaled-up gc for the observed stem ranged between 4.6 and 22.7 mm s−1 and predicts substantial reductions in canopy conductance when stem drops because of water stress (Fig. 5). In the energy balance model, LE ranged from 250 to 800 W m−2 (Fig. 6). The LE showed a crossover point between curves modeled for different ga values. In general, the implication of this shape of the curves is that a low wind speed day (low ga ) will have a lower ET at high canopy conductance (wet soil, high stem, open stomata) but a higher ET for low canopy conductance (wet soil, high stem, open stomata) compared to a high wind speed day (high ga ). The model for unstressed LE showed a reasonable agreement with the observed LE measured by the flux tower (Fig. 7). Unstressed LE ranged from 300 to 850 W m−2 and generally was higher and presented more noise than the observed LE. Peaks in the unstressed LE are associated with days having high aerodynamic conductance (not shown). The modeled LE expected for the average stem of the orchard failed to show a consistent reduction in days of stress with respect to the observed LE due to the variability of observed LE. However, the modeled stressed LE was always lower than the modeled unstressed LE during days of stress. Aerodynamic conductance during days of water stress was comparable to the aerodynamic conductance during days of well watered conditions, ranging from 7 to 53 mm s−1 and from 15 to 41 mm s−1 , respectively (Fig. 7).

When the proportional reduction of the stressed LE with respect to unstressed LE was used to estimate the expected reduction in Ka , clear expected reductions in Ka were associated with periods of water stress, with Ka reducing to 0.65 and 0.72 (Fig. 8). 4. Discussion 4.1. The disconnect between individual leaf and canopy response to stress The relationship between stomatal conductance and water stress measured on individual leaves contrasts with canopy conductance estimates made with the “big leaf” model (Monteith, 1965, 1981) equation using LE measured as the residual of the energy balance with H from eddy covariance. It may be argued that stomatal conductance is just one of three variables that determine leaf transpiration, i.e., the others being leaf temperature, vapor pressure at the leaf surface, leaf area etc. Leaf transpiration is reduced by stomatal closure, but it seems, not as much as stomatal conductance alone would imply (Lambers et al., 2008, p. 53), because transpiration is not a linear function of stomatal conductance. To obtain a more robust estimate of leaf transpiration, local environmental conditions in the canopy must be considered. Unfortunately, humidity data for the project came from the nearby CIMIS

18

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

Fig. 5. Seasonal pattern of the unstressed baseline for stem water potential (circle) and of the average stem water potential measured in the field (dot). Seasonal pattern of the unstressed canopy conductance (empty square) calculated scaling up the stomatal conductance expected from baseline and seasonal pattern of canopy conductance of the orchard (filled square) calculated scaling up the stomatal conductance expected from the average stem water potential of the orchard. The unstressed surface conductance for grass from FAO56 is indicated by the dashed horizontal line as a reference. The date of harvest is indicated by the dotted vertical line.

Fig. 6. Theoretical behavior of latent heat flux as a function of canopy conductance for three levels of aerodynamic conductance (ga ), and the following environmental conditions: absorbed radiation = 1100 W/m2; air vapor pressure = 1400 Pa; air temperature = 27 ◦ C.

station rather than within and above the almond orchard. Thus, the humidity data used in the big leaf model calculations may not represent the actual humidity within or above the orchard. Even with local humidity data, local differences in water status among trees may have created local differences in environmental variables. The flat response of measured Ka over time indicates a lack of water stress in the trees. The reduced photosynthesis corresponding to decreased stem suggests that plants are undergoing water stress, at least at the leaf level. Since photosynthetic rate should ultimately determine biomass assimilation and hence productivity, the recorded water stress likely reduced yield.

Our data on leaves shows that water stress causes stomatal closure, but how much the reduction in stomatal conductance aggregates to canopy conductance is a function of leaf area and of the population distribution of sunlit and shaded leaves in a canopy. Scale-up models indicate that stomatal closure predicts a substantial decrease in gc . Shaded leaves respond less dramatically to water stress and the fraction of shaded leaves in the canopy is larger than for sunlit leaves. It could be argued that when water stress occurs, sunlit leaves reduce stomatal conductance but the aggregate gc does not respond to water stress because of the overriding contribution from shaded leaves. However, the contribution to total gc by shaded and sunlit leaves, weighted by the relative LAI, was about

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

19

Fig. 7. Seasonal pattern of observed LE (empty diamond), modeled LE for an unstressed canopy (empty square) and modeled LE for the measured average stem water potential of the orchard (filled square). Seasonal pattern of midday stem water potential (filled dot) and baseline for stem water potential (empty dot). All values are averages at midday (12:00–15:00). Values in parentheses are average midday aerodynamic conductance in mm/s.

Fig. 8. Expected reduction (square) in the observed Ka (diamond) based on the proportional reduction of modeled stressed ET compared to modeled unstressed ET (Ks ). Stem water potential and baseline (black dot and circle) are indicated as a reference.

60:40, and this resulted in a substantial reduction of aggregate gc with stem . Stomatal control of transpiration is often overestimated (Hsiao, 1990) because much evidence for stomatal control of transpiration comes from experiments conducted in leaf chambers, where the conditions are very different from the field; in particular, aerodynamic conductance is increased by stirring air with fans. To obtain the expected response of LE to stomatal closure in the field, we modeled the energy balance using the measured ga and the gc obtained from the scale-up models.

4.2. The expected decrease in canopy ET with stress The energy balance model predicted that a lower LE would be expected during periods of low  stem due to reductions in gc (Fig. 5), but such a drop in LE was not observed, either on an absolute basis or with respect to measured canopy LE (Fig. 7). In theory, the unresponsiveness of LE to reduced gc could be due to the influence of aerodynamic conductance, combined with the predicted flat response of LE to gc at low ga (Jarvis, in Grace, 1981). However, both Jarvis’s and our models predict a substantially reduced LE (half) for a five to 23 mm s−1 reduction in gc and our observed average ∼50 mm s−1 ga . Our measured ga corresponded to that expected

20

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

for tall trees and predicted coupled conditions, but the insensitivity of observed LE to gc (decoupled response) was not predicted by the models. The same insensitivity of LE to gc was reported in threemeter maize with gc 20 to 40 mm s−1 , which is comparable to our ga for water vapor 20 to 75 mm s−1 . Note that the FAO reference surface has ga of 0 to 25 mm s−1 (Steduto and Hsiao 1998a). In that system, LE did not decrease until gc dropped below a mean threshold of 15 mm s−1 , although Jarvis’ curves predict that LE should drop ∼20–23% prior to reaching this threshold. The response of LE to water stress gets more complicated when the measured ETa is divided by ETo to address environmental variability. The reference ET surface has a very different aerodynamic conductance than tall canopy trees, which introduces noise that makes the response to stress more difficult to resolve. The Ka values are particularly unstable at high gc and for tall crops when compared with ETo (Annandale and Stockle, 1994). Both conditions are met by the wellwatered almond orchard, hence it may be appropriate to use an almond-specific reference ET (see below).

4.3. Water use efficiency and water stress It is generally assumed that intrinsic WUE will increase when stomata close, a conclusion drawn from analyzing which factors will increase the ratio of the leaf gas exchange equations for photosynthetic rate and transpiration rate (Jones, 1992[page 280, Fig. 10.6, p. 283], Steduto et al., 2007). The same approach is used in carbon isotope discrimination (Farquhar et al., 1989) when the ratio Ci /Ca is used as a first approximation of WUE. Our leaf gas exchange data show increased leaf WUE with water stress, consistent with previous reports. Similarly, the decrease in assimilation as stomatal conductance decreases in an A/Ci curve can be considered (Lambers et al., 2008, p. 53). The curved behavior of assimilation makes the ratio assimilation/stomatal conductance increase as stomata close. However, these considerations are only valid for intrinsic WUE (the ratio between photosynthetic rate and stomatal conductance) and cannot be extended to observed WUE (defined as photosynthetic rate/transpiration rate) because the relationship of stomatal conductance to transpiration is not linear, it is based on the mathematical relationships of leaf gas exchange equations, and it doesn’t consider the feedback effect of leaf temperature on the fluxes of carbon into and water vapor of out of the leaf as stomata close (Seibt et al., 2008).

4.4. Leaf temperature and water use efficiency Our study showed ET insensitive to water stress, but it is reasonable to assume that canopy carbon assimilation decreased during periods of water stress. In this scenario, plant water stress would decrease WUE. A similar conclusion was suggested by Hsiao (in Jackson et al., 1993) based on theoretical considerations. He argued that if the ratio Ci /Ca remains constant in well-watered and stressed conditions (Wong et al., 1979), the gradient for carbon assimilation should not change with stress. Conversely, the increase in leaf temperature caused by a stomatal closure increases the saturated vapor pressure in the leaf and thus the gradient for transpiration. This is a feedback effect that makes the response of transpiration relatively insensitive to small decreases in gs . At the same time, the increased temperature in summer crops in warm areas is likely to decrease the photosynthesis rate because photorespiration increases more than photosynthesis (Long et al., 2006). Therefore, the decrease in photosynthesis should be more than that predicted just by stomatal limitations. As a result, the real canopy WUE (canopy carbon flux/canopy ET) may not increase while stomata close or may even decrease.

4.5. Water use efficiency and coupling Another feedback mechanism for transpiration affects WUE. Under field conditions, transpiration from a leaf humidifies its environment, decreasing the gradient for transpiration, so that transpiration flux itself decreases the gradient for transpiration. Low coupling can also decrease the canopy assimilation rate, because the CO2 concentration at the leaf surface will decrease below that of the mixed layer above the canopy (Jarvis, 1985). However, Steduto and Hsiao (1998b) proposed that the influence of transpiration on air water vapor is much larger than the influence of assimilation on air CO2 concentration. Transpiration is limited mainly by the energy supply to vaporize water and secondarily by stomatal resistance to vapor transfer, while assimilation is mostly limited by stomatal conductance, mesophyll conductance, and the rate at which chloroplasts fix carbon. Under low coupling conditions, water stress may decrease assimilation more than transpiration, decreasing WUE (Baldocchi et al., 1983). The decreased assimilation may be caused by both stomatal and non-stomatal consequences of water stress. Hence, WUE may decrease with water stress due to non-stomatal limitations to photosynthesis that reduce the demand for CO2 (Jarvis, 1985). 4.6. Implications for water management The implications of the insensitivity of ET to water stress for the water budget of the State of California is that almonds orchards may need to be severely stressed (conceivably resulting in important loss in yields) before irrigation water is saved in almonds. The observed range of per-acre yield across the State spans from 2500 to 5000 lb., while the observed ET present a smaller range (midseason Kc ranges from 1.15 to 1.20, Fig. 2 and unpublished data) suggesting that the best efficiency in terms of ET/kernel yield may be found at the high end. However, measurement of canopy CO2 flux with eddy covariance is needed to confirm the hypothesis of a decreasing canopy-level WUE with water stress. 4.7. Measuring water stress from a decrease in ET An additional implication of our study is that it may be difficult to quantify water stress in almond using techniques that measure field ET, either with a field-based approach (eddy covariance, surface renewal) or with satellite imagery, because the response of LE to gs depends on ga . On a high ga (windy) day it will be more likely to measure a stress-related reduction in ET than during a low ga (calm) day. The task is more difficult if the short canopy (i.e. grass) ETo is used to account for the environment, since the inherent fluctuation of the ratio ETa /ETo (Annandale and Stockle, 1994) may add enough noise to lose the signal of plant water stress. In this situation, an unstressed, crop-specific reference ET could be calculated if functions were developed to estimate the crop ga as a function of wind speed. However, no such functions have been proposed for almonds or other crops. Indeed, such a crop-specific reference ET would pose the challenge of developing corrections for biological effects such as leaf out date, canopy development, senescence etc. Additionally, a reference ET for an unstressed tall, irrigated crop, will have large values for both ga and gc . Under our environmental conditions, large ga and gc will cause a highly variable LE, given the modeled relationship of LE to gc and ga (Fig. 6) resulting in a highly variable reference. Another challenge of measuring plant water stress using the ETa /ETo ratio is posed by the effect of LAI. In agricultural crops, even when there is water-stress-induced stomatal closure, gc (and thus ETa ) can be relatively insensitive because the reduction in gs may be overridden by the high LAI (Steduto and Hsiao, 1998b). Interestingly, in almond, LAI is lower at the beginning and end of the season. During these periods, we expect gc to

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

be more dependent on gs and it is more likely that stress can be measured using a reduction in ETa . However, during these periods the weather generally has a low evaporative demand and thus the ratio ETa /ETo can be highly variable, making it more difficult to pick up a stress-induced signal. Therefore, the idea of using a simple ratio of ETa /ETo to account for the environment, although very practical, may be too simplistic. Taking the ratio of the variables in the Penman-Monteith equation for ETa and ETo , one obtains a collection of variables that are difficult to interpret and that still depend on ga (Eq. (9)). For all these reasons, the best method to obtain a physiological measure of stress from a field measurement of ET is to reverse the Penman-Monteith equation and solve for gc ; that is, the part of the transport path under physiological control. This requires measuring or estimating ga . If the unstressed gc and ga for different crops were made available to growers, they could then calculate the unstressed ETc for a specific crop. The problem, however, with this approach is determining the respective contributions of gs and LAI, as two very significant factors impacting gc . Thus, research on the developmental pattern of LAI during the season and on the relationship between LAI and gs in determining gc is also necessary. We conclude from this study that inducing mild to moderate water stress in almond orchards may not result in substantial water savings. Additionally, the relative insensitivity of canopy ET to stomatal closure suggests that irrigation scheduling methods based on in-field measurements of ET may fail to detect plant water stress. Finally, this work underscores the multiple effects of plant water stress in impacting the estimation of canopy-level water use efficiency. Increasing canopy-level water use efficiency should be a major objective of irrigation management and used as a litmus test to drive policy decisions on the allocation of water, particularly in periods of drought. Acknowledgements The authors wish to express their gratitude to the Plant Sciences Department and the Horticulture and Agronomy Graduate Group at UC Davis for the financial support. Supplies and equipment for conducting the study were funded by Joseph M. Ogawa Research & Teaching Endowment Award, UC Davis Henry A. Jastro Graduate Research Award and John & Terry Kubota Graduate Scholarship. The authors also wish to thank Paramount Farming Company and the Almond Board of California for the support in the data collection of this study. The authors are grateful to Sebastiano Di Martino and Mireia Corell for the excellent assistance in performing gas exchange measurements. References Alfieri, J.G., Kustas, W.P., Prueger, J.H., Hipps, L.E., Evett, S.R., Basara, J.B., Neale, C.M., French, A.N., Colaizzi, P., Agam, N., 2012. On the discrepancy between eddy covariance and lysimetry-based surface flux measurements under strongly advective conditions. Adv. Water Resour. 50, 62–78. Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements. Food and Agriculture Organization of the United Nations, Rome. Allen, R.G., Pruitt, W.O., Wright, J.L., Howell, T.A., Ventura, F., Snyder, R., Itenfisu, D., Steduto, P., Berengena, J., Baselga Yrisarry, J., Smith, M., Pereira, L.S., Raes, D., Perrier, A., Alves, I., Walter, I., Elliott, R.A., 2006. Recommendation on standardized surface resistance for hourly calculation of reference ETo by the FAO56Penman–Monteith method. Agric. Water Manage. 81 (1–2), 1–22. Annandale, J.G., Stockle, C.O., 1994. Fluctuation of crop evapotranspiration coefficients with weather—a sensitivity analysis. Irrig. Sci. 15, 1–7. Baldocchi, D.D., Luxmoore, R.J., Hatfield, J.L., 1991. Discerning the forest from the trees—an essay on scaling canopy stomatal conductance. Agric. For. Meteorol. 54, 197–226. Baldocchi, D.D., Verma, S.B., Rosenberg, N.J., Blad, B.L., Garay, A., Specht, J.E., 1983. Influence of water stress on the diurnal exchange of mass and energy between the atmosphere and a soybean canopy1. Agron. J. 75, 543–548. Bange, G.G.J., 1953. On the quantitative explanation of stomatal transpiration. Acta Bot. Neerlandica 2, 255–297.

21

Bradford, K., Hsiao, T., 1982. Physiological responses to moderate water stress. In: Physiological Plant Ecology II. Springer, pp. 263–324. Buck, A.L., 1981. New equations for computing vapor-pressure and enhancement factor. J. Appl. Meteorol. 20, 1527–1532. G. Burba, (2013) Eddy covariance method for scientific, industrial, agricultural, and regulatory applications: a field book on measuring ecosystem gas exchange and areal emission rates. Campbell, G.S., Norman, J.M., 1998. An Introduction to Environmental Biophysics, 2nd ed. Springer, New York. Castellví, F., Snyder, R., 2010. A comparison between latent heat fluxes over grass using a weighing lysimeter and surface renewal analysis. J. Hydrol. 381, 213–220. Chalmers, D.J., Burge, G., Jerie, P.H., Mitchell, P.D., 1986. The mechanism of regulation of bartlett pear fruit and vegetative growth by irrigation withholding and regulated deficit irrigation. J. Am. Soc. Hortic. Sci. 111, 904–907. Chalmers, D.J., Mitchell, P.D., Vanheek, L., 1981. Control of peach-tree growth and productivity by regulated water-supply, tree density, and summer pruning. J. Am. Soc. Hortic. Sci. 106, 307–312. Doorenbos, J., Pruitt, W., 1977. Crop water requeriments. In: Irrigation and Drainage Paper 24. FAO, Rome, pp. 179. Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989. Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant Biol. 40, 503–537. Fereres, E., Soriano, M.A., 2007. Deficit irrigation for reducing agricultural water use. J. Exp. Bot. 58, 147–159. Girona, J., Mata, M., Marsal, J., 2005. Regulated deficit irrigation during the kernel-filling period and optimal irrigation rates in almond. Agric. Water Manage. 75, 152–167. Goldhamer, D.A., Viveros, M., Salinas, M., 2006. Regulated deficit irrigation in almonds: effects of variations in applied water and stress timing on yield and yield components. Irrig. Sci. 24, 101–114. Grace, J., Ford, E.D., Jarvis, P.G., 1981. Plants and Their Atmospheric Environment: The 21st Symposium of the British Ecological Society, Edinburgh, 1979. Blackwell Scientific Publications. Hsiao, T.C., 1973. Plant responses to water stress. Annu. Rev. Plant Physiol. 24, 519–570. Hsiao, T.C., 1990. Plant-atmosphere interactions, evapotranspiration, and irrigation scheduling. Acta Hortic. 278, 55–66. Jackson, M.B., Black, C.R., North Atlantic Treaty, O., Scientific Affairs, D., Climate, N.A.R.W.o.I.S.o.P.i.a.C., 1993. Interacting Stresses on Plants in a Changing Climate. Springer-Verlag, Berlin; New York. Jarvis, P., 1984. Coupling of transpiration to the atmosphere in horticultural crops: the omega factor. I International Symposium on Water Relations in Fruit Crops 171, 187–206. Jarvis, P.G., 1985. Coupling of transpiration to the atmosphere in horticultural crops: the omega factor. Acta Hortic. 171, 187–205. Jarvis, P.G., McNaughton, K.G., 1986. Stomatal control of transpiration: scaling up from leaf to region. Adv. Ecol. Res. 15 (1), p.49. Jones, H.G., 1992. Plants and Microclimate: a Quantitative Approach to Environmental Plant Physiology. University Press, Cambridge. Kyaw Tha Paw, U., Meyers, T.P., 1989. Investigations with a higher-order canopy turbulence model into mean source-sink levels and bulk canopy resistances. Agric. For. Meteorol. 47, 259–271. Kyaw, T.P.U., 1987. Mathematical analysis of the operative temperature and energy budget. J. Therm. Biol. 12, 227–233. Lambers, H., Chapin, I.F.S., Chapin, F.S., Pons, T.L., 2008. Plant Physiological Ecology. Springer. Long, S.P., Zhu, X.G., Naidu, S.L., Ort, D.R., 2006. Can improvement in photosynthesis increase crop yields? Plant Cell Environ. 29, 315–330. McCutchan, H., Shackel, K.A., 1992. Stem–water potential as a sensitive indicator of water-stress in prune trees (Prunus-domestica L. cv French). J. Am. Soc. Hortic. Science 117, 607–611. McDermitt, D.K., 1990. Sources of error in the estimation of stomatal conductance and transpiration from porometer data. Hortscience 25, 1538–1548. Monteith, J., 1965. Evaporation and environment. Symp. Soc. Exp. Biol., 4. Monteith, J.L., 1981. Evaporation and surface temperature. Q. J. R. Meteorol. Soc. 107, 1–27. Monteith, J.L., Unsworth, H., 1990. Principles of Environmental Physics. Edward Arnold. Moreno, F., Fernández, J.E., Clothier, B.E., Green, S.R., 1996. Transpiration and root water uptake by olive trees. Plant Soil 184 (1), 85–96. Naor, A., Peres, M., Greenblat, Y., Doron, I., Gal, Y., Stern, R.A., 2000. Irrigation and crop load interactions in relation to pear yield and fruit-size distribution. J. Hortic. Sci. Biotechnol. 75, 555–561. Paw, K.T., Gao, W.G., 1988. Applications of solutions to non-linear energy budget equations. Agric. For. Meteorol. 43, 121–145. Romero, P., Botia, P., Garcia, F., 2004. Effects of regulated deficit irrigation under subsurface drip irrigation conditions on vegetative development and yield of mature almond trees. Plant Soil 260, 169–181. Rouhi, V., Samson, R., Lemeur, R., Van Damme, P., 2007. Photosynthetic gas exchange characteristics in three different almond species during drought stress and subsequent recovery. Environ. Exp. Bot. 59, 117–129. Schellenberg, D.L., Alsina, M.M., Muhammad, S., Stockert, C.M., Wolff, M.W., Sanden, B.L., Brown, P.H., Smart, D.R., 2012. Yield-scaled global warming potential from N2O emissions and CH4 oxidation for almond (Prunus dulcis)

22

G.M. Spinelli et al. / Agricultural Water Management 168 (2016) 11–22

irrigated with nitrogen fertilizers on arid land. Agric. Ecosyst. Environ. 155, 7–15. Seibt, U., Rajabi, A., Griffiths, H., Berry, J.A., 2008. Carbon isotopes and water use efficiency: sense and sensitivity. Oecologia 155, 441–454. Shackel, K., 2011. A plant-based approach to deficit irrigation in trees and vines. Hortscience 46, 173–177. Shackel, K., Ahmadi, H., Biasi, W., Buchner, R., Goldhamer, D., Gurusinghe, S., Hasey, J., Kester, D., Krueger, B., Lampinen, B., McGourty, G., Micke, W., Mitcham, E., Olson, B., Pelletrau, K., Philips, H., Ramos, D., Schwankl, L., Sibbett, S., Snyder, R., Southwick, S., Stevenson, M., Thorpe, M., Weinbaum, S., Yeager, J., 1997. Plant water status as an index of irrigation need in deciduous fruit trees. HortTechnology 7, 23–29. Shapland, T.M., McElrone, A.J., Paw, U.K.T., Snyder, R.L., 2013. A turnkey data logger program for field-scale energy flux density measurements using eddy covariance and surface renewal. Ital. J. Agrometeorol. Sinclair, T.R., Murphy, C.E., Knoerr, K.R., 1976. Development and evaluation of simplified models for simulating canopy photosynthesis and transpiration. J. Appl. Ecol. 13, 813–829. Snyder, R.L., Pruitt, W.O., 1992. Evapotranspiration data management in California. In: Irrigation and Drainage Session, Proceedings Water Forum 1992, ASCE, August 2–6, 1992, Baltimore, MD, pp. 128–133. Steduto, P., Hsiao, T.C., 1998a. 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.

Steduto, P., Hsiao, T.C., 1998b. Maize canopies under two soil water regimes—III: Variation in coupling with the atmosphere and the role of leaf area index. Agric. For. Meteorol. 89, 201–213. Steduto, P., Hsiao, T.C., Fereres, E., 2007. On the conservative behavior of biomass water productivity. Irrig. Sci. 25, 189–207. Tanner, C.B., Sinclair, T.R., 1983. Efficient Water Use in Crop Production: Research or Re-Search?1. In: Taylor, H.M., Wayne, J.R., Thomas, S.R. (Eds.), Limitations to Efficient Water Use in Crop Production. American Society of Agronomy, Crop Science Society of America, Soil Science Society of America, pp. 1–27. Teviotdale, B., Goldhamer, D., Viveros, M., 2001. Effects of deficit irrigation on hull rot disease of almond trees caused by Monilinia fructicola and Rhizopus stolonifer. Plant Dis. 85, 399–403. Todorovic, M., 1999. Single-layer evapotranspiration model with variable canopy resistance. J. Irrig. Drain. Eng.-ASCE 125, 235–245. Torrecillas, A., Alarcon, J.J., Domingo, R., Planes, J., SanchezBlanco, M.J., 1996. Strategies for drought resistance in leaves of two almond cultivars. Plant Sci. 118, 135–143. Wong, S., Cowan, I., Farquhar, G., 1979. Stomatal conductance correlates with photosynthetic capacity. Zarate-Valdez, J.L., Whiting, M.L., Lampinen, B.D., Metcalf, S., Ustin, S.L., Brown, P.H., 2012. Prediction of leaf area index in almonds by vegetation indexes. Comp. Electron. Agric. 85, 24–32.