The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the official position of the American Society of Agricultural and Biological Engineers (ASABE), and its printing and distribution does not constitute an endorsement of views which may be expressed. Technical presentations are not subject to the formal peer review process by ASABE editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an ASABE meeting paper. EXAMPLE: Author's Last Name, Initials. 2011. Title of Presentation. ASABE Paper No. 11----. St. Joseph,Mich.: ASABE. For information about securing permission to reprint or reproduce a technical presentation, please contact ASABE at rutter@asabe.org or 269-932-7004 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).

An ASABE Meeting Presentation

Paper Number: 1111555

Plant Water Stress Detection Using Leaf Temperature and Microclimatic Information

Vasu Udompetaikul Department of Biological and Agricultural Engineering University of California, Davis, California

Shrini K. Upadhyaya Department of Biological and Agricultural Engineering University of California, Davis, California

David Slaughter Department of Biological and Agricultural Engineering University of California, Davis, California

Bruce Lampinen Department of Plant Sciences University of California, Davis, California

Ken Shackel Department of Plant Sciences University of California, Davis, California

Written for presentation at the 2011 ASABE Annual International Meeting

Sponsored by ASABE Galt House

Louisville, Kentucky August 7 – 10, 2011

Abstract. A proximal sensor suite consisting of an infrared thermometer, an air temperature sensor, a humidity sensor, a PAR sensor, and an anemometer was developed to measured leaf temperature and other relevant microclimatic information to determine plant water status. A series of experiments were conducted in almond and walnut orchards to study relationship between data obtained using the sensor suite and stem water potential measured using a standard pressure chamber. Multiple linear regression models of leaf temperature as functions of stem water potential, air temperature, relative humidity, photosynthetically active radiation, and wind speed were developed and validated for almond and walnut crops under sunlit and shaded conditions. Models yielded high correlation

The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the official position of the American Society of Agricultural and Biological Engineers (ASABE), and its printing and distribution does not constitute an endorsement of views which may be expressed. Technical presentations are not subject to the formal peer review process by ASABE editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an ASABE meeting paper. EXAMPLE: Author's Last Name, Initials. 2011. Title of Presentation. ASABE Paper No. 11----. St. Joseph,Mich.: ASABE. For information about securing permission to reprint or reproduce a technical presentation, please contact ASABE atrutter@asabe.org or 269-932-7004 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).

with R2 values ranging from 0.82 to 0.90. Discriminant analyses of the data obtained from the sensor suite resulted in error rates of 9 to 11% in walnuts and 16 to 17% in almonds. However, critically wrong decision error, which is the overall misclassification of stressed trees, was limited to 5 to 10% in almonds, and 2 to 7% in walnuts. Since shaded leaf datasets were better correlated to plant water status in regression analysis and resulted in good discrimination power in classification analyses, shaded leaf data that is easier to gather using the sensor suite may be used in future studies.

Keywords. plant water stress, plant water status, leaf temperature, infrared thermometer, discrimination analysis.

2

Introduction California is the nation’s sole commercial producer of almonds and walnuts (CDFA, 2009). In 2007, more than 1 million tons of almonds and walnuts were produced in California amounting to more than $3 billion in value. Goldhamer (1996, 1998) estimated that almond and walnut productions require approximately 2.4 and 4.2 m3, respectively, of water for each kilogram of nut. Therefore, California needs approximately 3 billion m3 of water to produce almonds and walnuts. Because water resource is becoming scarce and urban water demand is increasing, there is an urgent need to utilize water wisely for agricultural production. The key is to develop irrigation strategies for a better water use efficiency without affecting quality or quantity of yield. This requires monitoring water status of the plant frequently to properly manage irrigation.

Pressure chamber has been used widely to measure leaf or stem water potential for plant water status determination and irrigation scheduling for many crops (Chauvin et al., 2006; Lampinen et al., 2001; Naor, 2000; Shackel et al., 1997). However, this conventional method is tedious and time consuming, and frequently result in an inadequate amount of sampling (Cohen et al., 2005), and is not suitable for commercial applications (Jones, 2004).

To address these concerns, techniques based on measuring canopy temperature have been developed. When a plant is under stress due to lack of water, it tends to close the stomata to decrease transpiration leading to an increase in leaf temperature. The energy balance of a leaf shows that this change in leaf temperature also depends on ambient conditions (i.e., relative humidity, wind speed, and ambient temperature) and radiation incident on the canopy surface. Fortunately, these parameters can be easily measured in real-time using commercially available sensors. Sensing canopy temperature using infrared thermometers or thermal cameras has shown good potential to estimate plant water status for irrigation scheduling in cotton, corn, grapevine, and pistachios (González-Dugo et al., 2006; Moller et al., 2007; Payero and Irmak, 2006; Testi et al., 2008). Thermal imaging technique can be scaled up to large areas of crop (Jones, 2004) but involves image processing techniques and can be expensive. A simple infrared thermometer with proper acquisition techniques could be used as a rapid and noncontact sensing device to evaluate plant water status.

The objective of this study was to study the relationship between plant water stress, leaf temperature, and microclimatic parameters and develop classification tools based on plant leaf temperature to discriminate plant water status in almond and walnut trees.

Thermal sensing for plant water status Response of a plant leaf to plant water status and environmental parameters can be presented by an energy balance scheme (Jackson et al., 1988; Jones, 1992) which mainly consists of net radiation, sensible heat mostly by convection, and latent energy by evaporation across the leaf surface. For a leaf, energy generated form metabolic processes can be neglected. This model shows that leaf temperature depends on air temperature, relative humidity, solar radiation, leaf resistance, and boundary layer resistance. By utilizing proper sensors, we can study the relationship between these parameters. Air temperature, relative humidity, and solar radiation could be easily measured.

Leaf temperature can be measured remotely using an infrared thermometer (IRT) by detecting infrared energy emitted. Boundary layer resistance depends mainly on the shape and size of the leaf and wind speed. Wind passing through the leaves of many plants can be approximated as laminar flow over flat plates (Gates, 1980; Monteith and Unsworth, 2008). Leaf resistance (rL) can typically be measured using leaf porometer. Torrecillas et al. (1988) and Shackel

3

(2007) found good correlation between leaf and stem water potential and leaf resistance in almonds. In other words, plant water status affects to leaf resistance and temperature.

Materials and method

Sensor suite A sensor suite to measure leaf temperature and microclimatic information was developed to study the relationship between leaf temperature along with relevant microclimatic information and plant water status in almonds (Figure 1). It consists of an infrared thermometer (IRT) (Model 6000L, Everest Interscience, Tucson, AZ), a PAR sensor (LI-190, LICOR inc., Lincoln, NE), an air temperature and humidity probe (HMP35C, Visalia Inc., Woburn, MA), and an anemometer (WindSonic, Gill Instruments Ltd., Hampshire, UK).

Figure 1. Sensor suite consisting of an infrared thermometer, a PAR sensor, an air temperature

and humidity probe, and an anemometer. All sensors are interfaced with a datalogger. A pressure chamber was used to measure stem water potential. All instruments

were installed on a mobile cart to move through the orchard.

Experimental technique For walnuts, experiments were conducted in 2010 growing season in a Howard walnut (8 years old) and a Chandler walnut (4 years old) orchards located in Arbuckle, CA. For almonds, the experiments were directed in a 5-year-old orchard in Arbuckle, CA, and a 4-year-old orchard in Madera, CA. Both orchards were planted to nonpareil almond variety. In each orchard, 15 trees with various plant water deficit levels were observed several times to test the suitability of this sensor suite to determine plant water status. During a given test (visit to the orchard), leaf temperature, PAR, air temperature, RH, and wind speed were measured using the sensor suite on each tree during 1 to 4 PM at which plant experiences daily maximum water stress (Lampinen et al., 2004). SWP was measured using a pressure chamber. IR sensor was used

4

to measure the temperatures of five sunlit and five shaded leaves per tree. Each observation consisted of averages of 5 sample of leaf temperature (TL), air temperature (TA),photosynthetically active radiation (PAR), relative humidity (RH), and wind speed (vA)measurements. In addition, one stem water potential (SWP) measurement was taken per tree. This experimental procedure was repeated between 3 to 6 times for a given orchard to archive a wide range of ambient conditions.

AnalysisIn 2010, we developed models based on temperature difference between leaf and ambient air (TL-TA) as functions of SWP, PAR, vapor pressure deficit (VPD), and wind speed with high coefficient of multiple determination (R2 values) of 0.83 for almonds and 0.84 for walnuts (Udompetaikul et al., 2010). However, those models included both sunlit and shaded leaves and distribution of light data was essentially bimodal. Range of light levels intercepted by shaded leave was narrow and it was interesting to observe the shaded leaves separately. In addition, VPD is also function of air temperature and relative humidity. It may be appropriate to use air temperature and relative humidity as independent variables instead of a single variable, VPD. This suggests the use of leaf temperature rather than difference between the temperature of the leaf and air as the dependent variable.

Based on the above, the data were analyzed using SAS software package (SAS Institute, Inc. v.9.2. Cary, NC). Multiple linear regression analysis was used to study relationship between leaf temperature, plant water status, and microclimatic information. By utilizing stepwise model selection approach with k-fold cross validation (Hastie et al., 2009; SAS, 2010), empirical models for leaf temperature as functions of SWP, PAR, air temperature, RH, and wind speed were developed for each crop and light exposure conditions. Second order polynomial model was used to account for quadratic effects, if any.

Our ultimate interest is to predict plant water status using the data obtained from the sensor suite. However, the leaf temperature was a dependent variable in our model and the plant water stress was an independent variable. We propose a technique to classify the plant water status as stressed and unstressed trees based on the critical values of stem water potential. The prediction models were used to determine critical values of the leaf temperature ( )corresponding to critical values of stem water potential (SWPc). Plant will be classified as stressed if its leaf temperature, , is higher than . Classification accuracy was computed by comparing predicted stress to the measured stress level.

Actual tree stress level was defined by considering the plant water potential below the baseline, which is maximum SWP achieved when plant gets fully irrigated. This baseline depends on crop type and vapor pressure deficit. Baseline functions for almonds and walnuts are given by (McCutchan and Shackel, 1992; Shackel et al., 1997):

Baseline (bar) = –1.20 VPD – 4.10 for almonds,

= –0.64 VPD – 2.78 for walnuts.

The plant stress threshold was defined as a straight line parallel to the baseline. In our study, the plant stress threshold was placed under the baseline by 8 bars and 4 bars for almonds and walnuts, respectively. SWP value on the threshold line is the critical SWP (SWPc). A Tree was defined as stressed if the measured SWP is lower than the SWPc at that ambient condition (i.e., VPD value). This criterion was used to define the true stress level for the tree samples to evaluate the classification power of each discriminant analysis also.

Further studies using discriminant analyses were performed to distinguish plants into two groups, stressed and unstressed, from leaf temperature, air temperature, RH, PAR, and wind

5

speed data. Stepwise discriminant analysis was used to select a subset of the quantitative variables for use in discriminating trees among the two classes – stressed and unstressed (Klecka, 1980). Moreover, a canonical discriminant analysis was used to find canonical variables which are linear combinations of the quantitative variables that provide maximal separation between classes (SAS, 2010). Since only two classes were involved in this study, one canonical variable was necessary. Separate analyses were conducted for each crop and light exposure condition. Classification accuracies of discriminant models were determined by performing leave-one-out cross-validation technique (Khattree and Naik, 2000). All classification techniques were compared to suggest suitable models to discriminate between stressed and unstressed trees.

Results and discussions

Multiple regression models Based on multiple linear regression analysis of 193 and 74 observational data respectively from almond and walnut trees (table 1), strong relationships between leaf temperature and stem water potential with other microclimatic information were found. Since relationship between parameters was not necessarily linear, a second order polynomial model was used. Prediction models yielded high R2 values of 0.872 and 0.897 for sunlit and shaded leaves, respectively, for almonds, (Eqns. 1 and 2 in table 2). For walnuts, R2 values were 0.835 for sunlit leaves and 0.875 for shaded leaves (Eqns. 3 and 4). Both SWP and TA were highly significant in all models, which showed a good relationship between TL, TA, and SWP. For sunlit leaves, PAR level was also significant. Note that PAR was also significant in shaded walnut leaf. Since the parameters utilized in the regression analysis were standardized and PAR value for shaded was

Table 1. Means, standard deviation, and range of observational data collected from two almond and two walnut orchards and categorized by different light exposure condition.

Almonds Walnuts Parameter Statistic Sunlit Shaded Sunlit Shaded PAR Mean 1818.4 203.3 1763.4 178.8

(�mol s-1 m-2) SD 180.1 28.3 149.8 45.3 Range 1239 to 2131 135 to 284 1320 to 2047 78 to 249

TL Mean 33.8 29.0 38.4 25.4 (°C) SD 2.9 2.6 5.2 2.7

Range 27.6 to 42.4 22.2 to 34.2 28.9 to 48.1 18.2 to 31.8 TA Mean 30.0 30.7

(°C) SD 2.4 1.9 Range 23.3 to 33.8 25.3 to 33.8

RH Mean 40.5 31.8 (%) SD 5.6 5.0

Range 27.2 to 55.9 17.5 to 45.8 Wind speed Mean 0.52 0.44

(m s-1) SD 0.22 0.24 Range 0.16 to 1.30 0.08 to 1.42

SWP Mean -15.6 -7.8 (bar) SD 2.4 2.6

Range -22 to -10 -13.8 to -3.5 No. of observations 193 74

6

relatively low, the square term for the standardized PAR was even lower. When combined with a low regression coefficient (0.34), the effect from PAR was marginal for this model. Wind speed was significant in sunlit leaf models for both crops whereas the RH effect was significant in shaded walnut leaf model.

In addition, linear models, that contain no interaction and quadratic terms, were also explored because of their simplicity. Strong relationships were still found with slightly lower R2 values compared to the full quadratic models. For almonds, the models yielded high R2 values of 0.855 for sunlit leaves and 0.894 for shaded leaves (Eqns.5 and 6 in table 3). For walnuts, linear models yielded R2 values of 0.821 for sunlit leaves and 0.862 for shaded leaves (Eqn.7 and 8). Because of their simplicity to explain the relationship with similar R2 values to the full quadratic model, the linear parameter models (Eqns. 5 to 8) were used in further studies.

Table 2. Second order polynomial models for leaf temperature as functions of SWP, TA, RH, PAR, and vA under different light exposure condition for almonds and walnuts, where superscript * shows that parameters are standardized by subtracting the mean and then dividing by the standard deviation (Table 1). Crop Exposure Empirical prediction model R2

Almonds Sunlit TL = 30.96 + 2.31 � 1.22 + 2.76 � 0.46

� 0.64 � 0.48 2 � 1.13 + 0.17

0.872 (1)

Shaded TL = 28.84 + 2.51 � 1.53 � 0.15 2 0.897 (2)

Walnuts Sunlit TL = 39.06 + 4.22 + 0.97 + 0.65 + 0.64 2 + 0.46

0.835 (3)

Shaded TL = 25.11 + 1.81 + 1.56 + 0.34 2 + 0.29 0.875 (4)

Table 3. Linear models for leaf temperature as functions of SWP, TA, RH, PAR, and vA under different light exposure conditions for almonds and walnuts, where superscript * shows that parameters are standardized by subtracting the mean and then dividing by the standard deviation (Table 1).

Crop Exposure Empirical prediction model R2

Almonds Sunlit TL = 31.05 + 2.44 � 2.20 + 2.47 � 0.36 0.855 (5)

Shaded TL = 28.68 + 2.49 � 1.45 0.894 (6)

Walnuts Sunlit TL = 38.43 – 4.06 � 0.10 + 0.66 + 0.51 0.821 (7)

Shaded TL = 25.44 – 1.72 + 1.46 + 0.29 0.862 (8)

Plant water stress classification using regression models When critical SWP values were used in eqns. 5 and 6 for almonds, and 7and 8 for walnuts, critical leaf temperatures as stress thresholds could be calculated. If measured leaf temperature is higher than this critical temperature, the tree is classified as stressed. Table 4 shows result of misclassification from this classification analysis. This approach had total errors

7

of 10.8 and 12.3% for sunlit and shaded walnut trees, respectively. In terms of misclassification of stressed trees as unstressed trees, which is designated as “critically wrong decision” as this has implications on plant growth and yield, the error rates were 5.4% and 6.9% for sunlit and shaded walnut trees, respectively. A less serious error that would lead to over-irrigation (i.e., detecting unstressed trees as stressed) was about 5.5% for all walnut trees. However, in almonds, errors were higher, total errors were 17.6 and 15.0% for sunlit and shaded leaves, respectively. This technique resulted in 8.8 and 5.2% critically wrong decision, and 8.8 and 9.8% over-irrigation decision for sunlit and shaded leaves, respectively. Small size of almond leaves may have been affected by air temperature much more than in walnuts resulting in poorer results in almonds compared to walnuts. In spite higher level of errors in almond, the critical errors were still below 10%.

Table 4. Classification of almond and walnuts trees into stressed and unstressed trees based on regression models with critical SWP value under different light exposure condition.

Error rate (%) Overall error (%)Crop Exposure

Stressed Unstressed Total CriticalError

OverIrrigation

Almonds Sunlit 15.7 20.0 17.6 8.8 8.8 Shaded 9.4 21.8 15.0 5.2 9.8

Walnuts Sunlit 11.4 10.3 10.8 5.4 5.4 Shaded 14.7 10.3 12.3 6.9 5.5

Discriminant analyses Similar to previous classification analysis using MLR models, classification accuracies were evaluated from total error rate and critical error rate. Stepwise discriminant analysis resulted in total error rates for sunlit and shaded almond trees of 16.6 and 16.1%, respectively. The critically wrong decisions were 9.8% for sunlit and 7.3% for shaded conditions. In walnuts, total error rates were 10.8 and 9.6% for sunlit and shaded conditions, respectively. The critically wrong decision was 4.1% for both light exposure conditions. Leaf temperature, air temperature and humidity played important role in classifying almond trees. For sunlit leaves, PAR level was also important in classification analysis. Interestingly, sunlight level did not play a role in classifying walnut trees. For shaded walnuts trees, leaf and air temperature were major factors in assisting with the classification analysis.

Table 5. Discrimination accuracy of stepwise discriminant analysis under different light exposure condition in almonds and walnuts.

Error rate (%) Overall error (%) Crop Exposure

Stressed Unstressed Total CriticalError

OverIrrigation

Significantparameters

Almonds Sunlit 17.6 15.3 16.6 9.8 6.7 TL, TA, PAR, RH Shaded 13.2 19.5 16.1 7.3 8.8 TL, TA, RH

Walnuts Sunlit 8.6 12.8 10.8 4.1 6.8 TL, vA

Shaded 8.8 10.3 9.6 4.1 5.5 TL, TA

8

When the canonical discriminant analysis was used in almonds, total error rates were 16.1% for both light exposure conditions. In walnuts, error rates were 8.8% for sunlit and 9.6% for shaded conditions. In terms of critically wrong decision, there was 9.3% for sunlit and 7.8% for shaded leaves in almonds, and 2.0% for sunlit and 4.1% for shaded leaves in walnuts. This technique could discriminate stress levels in walnut trees using sunlit leaves very effectively by keeping the total and critical errors low.

Table 6. Discrimination accuracy for canonical discriminant analysis under different light exposure condition in almonds and walnuts.

Error rate (%) Overall error (%)Crop Exposure

Stressed Unstressed Total CriticalError

OverIrrigation

Almonds Sunlit 16.7 15.3 16.1 9.3 6.7 Shaded 14.2 18.4 16.1 7.8 8.3

Walnuts Sunlit 4.2 13.2 8.8 2.0 6.8 Shaded 8.8 10.3 9.6 4.1 5.5

Form the above discussion it is clear that MLR models show a better relationship between the plant water status and the leaf temperature for shaded leaves than sunlit leaves. For both the discriminant analyses, classification accuracies for sunlit and shaded leaves were not significantly different. However, amount of light interception normal to the leaf surface is necessary to make accurate classification of sunlit leaves. This is a very important and interesting outcome as sunlit leaves change leaf orientation depending on the light intensity making it tedious to obtain radiation data normal to the leaf surface. From a practical point of view, it is much more convenient to obtain shaded leaf data using the sensor suite. Results from this study suggest that good discriminant models for walnuts and almonds could be developed using only shaded leaf data.

ConclusionsA proximal sensor suite consisting of an infrared thermometer, an air temperature sensor, a humidity sensor, a PAR sensor, and an anemometer was developed to measured leaf temperature and other relevant microclimatic information to predict plant water status. Based on a series of experiments conducted in almond and walnut orchards to study relationship between data obtained from the sensor suite and the plant water status measured by a standard pressure chamber following conclusions can be drawn:

(i) Multiple linear regression models of leaf temperature as functions of stem water potential, air temperature, relative humidity, photosynthetically active radiation, and wind speed were developed and validated for almond and walnut crops under sunlit and shaded light exposure conditions. Models yielded high coefficient of multiple determination (R2) values from 0.82 to 0.90. By utilizing the concept of critical SWP value in regression models, critical leaf temperature were determined as the threshold value to discriminate tree as stressed or unstressed. This analysis yielded total error rates of 11 to 12% in walnuts and 27 to 38% in almonds. Stepwise discriminant analysis resulted in error rates of 10% to 11% in walnuts and about 16% to 17% in almonds. Canonical discriminant analysis resulted in error rates of 9% to 10% in walnuts and about 16% in almonds.

9

(ii) When critically wrong decision error rate, which is the overall misclassification of stressed tree, was considered, all the classification techniques yielded 9 to 10% in sunlit almonds, 5 to 8% in shaded almonds, 2 to 5% in sunlit walnuts, and 4 to 7% in shaded walnuts, respectively. The results showed that shaded leaf temperature yielded better correlation to plant water status compared to sunlit leaf temperature in regression analysis models and similar discriminant power in all classification analyses. These results suggest that only shaded leaves could be used in future studies.

AcknowledgementsThe authors would like to acknowledge National Institute of Food and Agriculture grant programs (SCRI-USDA-NIFA No. 2010-01213), Almond Board of California, California Walnut Board, and Henry A. Jastro Graduate Research Scholarship Award for the financial support to conduct these research activities. Moreover, we sincerely appreciate the Royal Thai Scholarship for providing financial support to Mr. Vasu Udompetaikul to pursue his graduate education at UC Davis.

ReferencesCDFA. 2009. California Agricultural Resource Directory 2008–2009. California Department of

Food and Agriculture. Sacramento, CA.

Chauvin, W., T. Ameglio, J. P. Prunet, and P. Soing. 2006. Irrigation of walnut trees managing the water potential. Acta Horticulturae 705:473-477.

Cohen, Y., V. Alchanatis, M. Meron, Y. Saranga, and J. Tsipris. 2005. Estimation of leaf water potential by thermal imagery and spatial analysis. J. Exp. Bot. 56(417):1843-1852.

Gates, D. M. 1980. Biophysical ecology. Springer-Verlag, New York.

Goldhamer, D. A. 1996. Irrigation Scheduling. In Almond production manual, p. 171-178. W. C. Micke, ed: Division of Agriculture and Natural Resources, University of California. .

Goldhamer, D. A. 1998. Irrigation Scheduling. In Walnut production manual, p. 159-166. D. E. Ramos, ed: Division of Agriculture and Natural Resources, University of California. Publication 3373.

González-Dugo, M., M. Moran, L. Mateos, and R. Bryant. 2006. Canopy temperature variability as an indicator of crop water stress severity. Irrigation Science 24(4):233-240.

Hastie, T., R. Tibshirani, and J. H. Friedman. 2009. The Elements of Statistical Learning. 2 ed. Springer.

Jackson, R. D., W. P. Kustas, and B. J. Choudhury. 1988. A reexamination of the crop water stress index. Irrigation Science 9(4):309-317.

Jones, H. G. 1992. Plants and microclimate : a quantitative approach to environmental plant physiology. 2nd ed. Cambridge University Press.

Jones, H. G. 2004. Irrigation scheduling: advantages and pitfalls of plant-based methods. J. Exp. Bot. 55(407):2427-2436.

Khattree, R., and D. N. Naik. 2000. Multivariate data reduction and discrimination with SAS software. SAS Institute Inc.

Klecka, W. R. 1980. Discriminant analysis. Sage University Paper Series on Quantitative Applications in the Social Sciences. Sage Publications, Beverly Hills and London.

10

Lampinen, B., K. Shackel, S. Southwick, and W. Olson. 2001. Deficit irrigation strategies using midday stem water potential in prune. Irrigation Science 20(2):47-54.

Lampinen, B. D., K. A. Shackel, S. M. Southwick, W. H. Olson, and T. M. DeJong. 2004. Leaf and canopy level photosynthetic responses of French prune (Prunus domestica L. 'French') to stem water potential based deficit irrigation. Journal of Horticultural Science and Biotechnology 79(4):638-644.

McCutchan, H., and K. A. Shackel. 1992. Stem-water Potential as a Sensitive Indicator of Water Stress in Prune Trees (Prunus domestica L. cv. French). J. Amer. Soc. Hort. Sci.117(4):607-611.

Moller, M., V. Alchanatis, Y. Cohen, M. Meron, J. Tsipris, A. Naor, V. Ostrovsky, M. Sprintsin, and S. Cohen. 2007. Use of thermal and visible imagery for estimating crop water status of irrigated grapevine. J. Exp. Bot. 58(4):827-838.

Monteith, J. L., and M. H. Unsworth. 2008. Principles of environmental physics. 3rd ed. Academic Press.

Naor, A. 2000. Midday stem water potential as a plant water stress indicator for irrigation scheduling in fruit trees. Acta Horticulturae 537:447-454.

Payero, J., and S. Irmak. 2006. Variable upper and lower crop water stress index baselines for corn and soybean. Irrigation Science 25(1):21-32.

SAS. 2010. SAS OnlineDoc, Version 9.2. Cary, N.C.: SAS Institute, Inc.

Shackel, K. A. 2007. Water relations of woody perennial plant species. Journal International des Sciences de la Vigne et du Vin 41(3):121-129.

Shackel, K. A., H. Ahmadi, W. Biasi, R. Buchner, D. Goldhamer, S. Gurusinghe, J. Hasey, D. Kester, B. Krueger, B. Lampinen, G. McGourty, W. Micke, E. Mitcham, B. Olson, K. Pelletrau, H. Philips, D. Ramos, L. Schwankl, S. Sibbett, R. Snyder, S. Southwick, M. Stevenson, M. Thorpe, S. Weinbaum, and J. Yeager. 1997. Plant water status as an index of irrigation need in deciduous fruit trees. HortTechnology 7(1):23-29.

Testi, L., D. Goldhamer, F. Iniesta, and M. Salinas. 2008. Crop water stress index is a sensitive water stress indicator in pistachio trees. Irrigation Science 26(5):395-405.

Torrecillas, A., M. Ruiz-Sanchez, A. Leon, and A. Garcia. 1988. Stomatal response to leaf water potential in almond trees under drip irrigated and non irrigated conditions. Plant and Soil112(1):151-153.

Udompetaikul, V., S. K. Upadhyaya, D. C. Slaughter, and B. D. Lampinen. 2010. Development of a sensor suite to determine plant water potential. Paper No: 1009450. 2010 ASABE Annual International Meeting. Pittsburgh, Pennsylvania, June 20 - June 23, 2010. St. Joseph, Mich.: ASABE.