AR

LHa

Pb

c

d

e

a

ARRAA

KMESPCH

1

mnci

Ebt(

C1

h0

Agricultural Water Management 140 (2014) 37–47

Contents lists available at ScienceDirect

Agricultural Water Management

jou rn al hom ep age: www.elsev ier .com/ locate /agwat

ssessing crop coefficients for Zea mays in the semi-arid Hailiutuiver catchment, northwest China

izhu Houa,b,∗, Jochen Wenningerc,d, Jiangen Shena, Yangxiao Zhouc,an Baoa, Haijun Liue

School of Water Resources and Environmental Science, China University of Geosciences (Beijing), No. 29, Xueyuan Road, Haidian District, Beijing 100083,R ChinaKey Laboratory of Groundwater Circulation and Evolution, China University of Geosciences (Beijing), Ministry of Education, Beijing 100083, PR ChinaUNESCO-IHE, Department of Water Science and Engineering, PO Box 3015, 2611 DA Delft, The NetherlandsDelft University of Technology, Water Resources Section, PO Box 5048, 2600 GA Delft, The NetherlandsCollege of Water Sciences, Beijing Normal University, Beijing 100875, PR China

r t i c l e i n f o

rticle history:eceived 18 August 2013eceived in revised form 20 March 2014ccepted 31 March 2014vailable online 19 April 2014

eywords:aize (Zea mays L.)

vapotranspirationap flow

a b s t r a c t

To improve irrigation water-use efficiency, plant transpiration and soil evaporation in a maize (Zea maysL.) field in the Bulang sub-catchment of the Hailiutu River catchment in Northwest China were determinedusing in situ measurements. Crop transpiration (Tp) rates from Jul 15 to Oct 1, 2011 were measured withsap flow sensors, and soil evaporation (Ep) rates were measured with micro-lysimeters under an absenceof water deficit. The two rates together gave the total evapotranspiration (ETc) of the maize field. Cumu-lative Tp and Ep were 245 and 85 mm, accounting for 74 and 26% of total ETc (330 mm), respectively. Tocalculate the total ETc rate of the maize field for the entire growing season, the Penman–Monteith equa-tion combined with a single crop coefficient method (FAO-56) was used. The estimated crop coefficient(Kc) was calibrated using actual sap flow and soil evaporation data to provide accurate estimates of actual

enman–Monteith equationrop coefficientailiutu River catchment

evapotranspiration. The total crop ETc of the maize field for the 2011 and 2012 growing seasons was 583and 500 mm, respectively, with a mean daily value of ∼4 mm d−1. Groundwater contributed 33% of themaize ETc in 2011 (average groundwater table of 1.12 m with full irrigation) and 27% in 2012 (averagegroundwater table of 0.89 m with full irrigation). These results will improve precise planning and efficientmanagement of irrigation for maize in this region.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

In Northwest China, and particularly the Hailiutu River catch-ent, natural water resources are limited and irrigation is a

ecessary agricultural practice. Therefore, accurate estimates ofrop evapotranspiration (ETc) are critical in order to assumenformed decisions regarding water management.

Two approaches can be used to estimate ETc by means ofT-based models or sensor-based measurements. The first com-

ines a reference crop evapotranspiration (ET0), which incorporateshe effects of various weather conditions, with a crop coefficientKc) that represents the entire potential evapotranspiration (ETp).

∗ Corresponding author at: School of Water Resources and Environmental Science,hina University of Geosciences, No. 29, Xueyuan Road, Haidian District, Beijing00083, PR China. Fax: +86 010 82321081.

E-mail address: houlizhu@gmail.com (L. Hou).

ttp://dx.doi.org/10.1016/j.agwat.2014.03.016378-3774/© 2014 Elsevier B.V. All rights reserved.

The impact of crop water stress is usually taken into account byusing a specific “stress function” allowing to determine the stresscoefficients, Ks, aimed to reduce potential (or maximum) cropevapotranspiration, ETp, (ETp = Kc ET0), in actual evapotranspiration,ETa, (ETa = Ks ETp), depending on the actual soil water content in theroot zone. Following a macroscopic approach, several models havebeen proposed for the stress function (Feddes et al., 1978; Homaeeet al., 2002; Li et al., 2006), as recently summarized by Rallo andProvenzano (2012). However, the applicability of the above methodrequires more extensive validation in arid and semi-arid regions,such as in the Hailiutu River catchment.

The second approach is the direct measurement of ETc using,for example, soil water budget or eddy covariance at daily toannual time scales (Wilson et al., 2001). Some measurement meth-

ods impact the natural environment, such as weighing lysimeters,which disturb the soil, or field chambers, which deduce ETc usingair humidity but also modify the microclimate (Kang et al., 2003;Chabot et al., 2005). Chemical and isotopic tracers (Robertson and

3 ter Ma

GcpBJltnrsboTohtatehTtuactb(

fiBdtsmtbMdcfepwR

2

2

oCcoe1aausca(ra

8 L. Hou et al. / Agricultural Wa

azis, 2006; Wenninger et al., 2010) have also been used to monitoranopy transpiration, although the data can be difficult to inter-ret, and these tracers do not permit sequential measurement. Theowen ratio method, the eddy correlation technique (Stockle and

ara, 1998; Droogers, 2000; Cammalleri et al., 2013), the scintil-ometric technique (De Bruin, 2002; Cammalleri et al., 2010), andhe aerodynamic combined method (Perrier and Tuzet, 1991) doot modify the natural environment and permit good temporalesolution. However, these methods are complex, require expen-ive equipment, and are affected by the uncertainty of partitioningetween crop transpiration (Tp) and soil evaporation (Ep). More-ver, for sparse vegetation, partitioning the measured fluxes intop and Ep remains a challenge, given that irrigation depends onlyn the former (Cammalleri et al., 2013). Sap flow measurementsave been used for several years to determine canopy transpira-ion (Swanson, 1994; Liu et al., 2008). These methods have severaldvantages, including the direct measurement of sap flux, rela-ive ease of use, continuous monitoring over time, and minimalnvironmental impacts. Several sap flow measurement methodsave been developed (Smith and Allen, 1996; Chabot et al., 2005).he energy balance method is one of the most established toolso directly measure transpiration, plant water stress, and waterse by plants (Green et al., 2006; Kigalu, 2007). These methodsre still under development, but they already show potential forontinuous measurements of crop water consumption with rela-ively inexpensive equipment. In addition, micro-lysimeters cane used successfully to monitor soil evaporation in rain fed fieldsFlumignan et al., 2011).

This paper describes an experimental study in an irrigated maizeeld located in the Hailiutu River catchment in Northwest China.ecause the evapotranspiration for the irrigated maize (ETp) wasetermined as the sum of potential transpiration (Tp) and poten-ial evaporation (Ep) using sap flow sensors (Dynamax-Flow 32ystem) and micro-lysimeters, respectively, the actual Kc measure-ents can be compared with the constructed Kc curves through

he FAO single coefficient method, where actual Kc was calculatedy dividing the measured ETp by ET0 following the FAO Penmanonteith equation. Through this comparison, the final Kc can be

etermined. The objectives of this study were: (1) to evaluate therop coefficients Kc during the middle and final phenological stagesor maize (Z. mays L.) in a semi-arid area of China; (2) to evaluate therrors in estimated ETc when using the average Kc values for bothhases of the phenological cycle; (3) to determine the total cropater use and the groundwater contribution to ETc in the Hailiutuiver catchment.

. Material and methods

.1. Experimental site

The experiment was carried out in the Bulang sub-catchmentf the Hailiutu River catchment, Shaanxi Province, Northwesthina (38◦23′33.87′′N, 109◦11′58.27′′E, 1282 m a.s.l; Fig. 1). Thelimate is semi-arid with a long-term average annual precipitationf 340 mm y−1 and an average annual potential evaporation (panvaporation) of 2184 mm y−1 (Wushenqi meteorological station,985–2004). Farmers in this area mainly practice subsistencegriculture, with irrigated maize as the primary crop. River waternd groundwater are used for irrigation in the valley, but in theplands, groundwater is the only option. The soil is classified asand according to the United States Department of Agriculturelassification, with average clay, silt, and sand contents of 0.2, 6.8,

nd 93.0%, respectively. The soil water contents at field capacitySWCfc) and wilting point (SWCwp) were 0.30 and 0.05 m3 m−3,espectively, determined from the average water retention curvelong the soil profile.

nagement 140 (2014) 37–47

Every 10 min from Apr 30 to Oct 1 in 2011 and from Jun 3 to Sep27 in 2012, the water table was measured using a Mini-Diver gauge(DI502, Eijkelkamp, Giesbeek, The Netherlands), and a Baro-Diver(DI500, Eijkelkamp) was used to measure the air pressure in agroundwater observation well located at the experimental site.The groundwater level range in inter-dune lowlands during themeasuring periods was 93.1–130.9 cm in 2011 and 55.6–116.6 cmin 2012.

2.2. Experimental layout and irrigation management

The pilot experiment covered ∼1.6 ha (145 × 110 m) croppedwith maize. The field was equipped with surface irrigation sys-tems. The experimental plot (1.8 × 9 m) was equipped with six sapflow sensors and had four rows of maize planted at 0.43 m spac-ing and an average density of 96 seeds, equivalent to ∼60,000plant/ha (Fig. 2). Because precipitation in the growing period was∼214 mm in 2011, much less than the total maize water require-ment of ∼583 mm, additional irrigation was required. The field wasirrigated with groundwater from a pumping well using a pipe sys-tem (Rain Bird, Azusa, USA). Irrigation volumes were measuredusing a stainless-steel triangle weir (RBC Flume, Eijkelkamp) witha flow range of 0.1–8.7 L s−1 (with a metrical accuracy of ±0.05%).The total irrigation water depth was 177 mm over six irrigationevents in 2011 and 60 mm over two irrigation events plus 411 mmof precipitation in the growing period of 2012. Irrigation eventsfollowed local management practices for the study area. In 2011,the irrigation water depths were 29.27, 29.37, 27.86, 29.9, 31.32,and 29.34 mm on May 25, Jun 21, Jun 30, Jul 15, Jul 26, and Aug 7,respectively. Both irrigations in 2012, on Jun 29 and Jul 13, were30 mm in depth.

2.3. Crop measurements

The main stem diameter 25–30 cm above the ground surface andnew shoot length were measured every month on nine plants in theexperimental plot. Plant height and leaf area index (LAI) were mon-itored 2–3 times a week throughout the growing season (Fig. 3).Plant height was measured by a tapeline from the soil surface to theplant apex (before heading) or to the crest of the spike (excludingawn, after heading). Leaf area (LA) was determined using leaf length(L), maximum width (W), and the formula LA = 0.7634 × L × W (Liuet al., 2011). LAI was calculated as:

LAI = LA9samles

9�crop (1)

where LA9 samples is the total leaf area of the nine plant samples (m2)and �crop is the plant density per unit area (m−2).

In the experimental field, the average maize yields were 10,527and 13,703 kg ha−1, and at harvest, plants were 262 cm tall in 2011and 293 cm tall in 2012. These yields and heights were similar tothe data reported by Du (2011) for maize in Yuyang District, YulinCity, Shaanxi Province, China.

2.4. Sap flow and soil evaporation

ETc under an absence of water deficit generally consists of Tp

and Ep. In this study, Tp was measured using sap flow sensors andEp with micro-lysimeters. Stem flow gauges (Flow32; Dynamax,Houston, TX, USA) were used with the energy balance method tomeasure sap flow (SF) in the maize stems. When the plants were∼180 cm high with a stem diameter of ∼25 mm (76 days after sow-

ing), six plants were selected for SF measurements (Fig. 2). Flow32sensors (110 mm long, 24–32 mm diameter) were installed on theirstems at a height of 25–30 cm above the ground on Jul 15, 2011,when the plants were fully developed.

L. Hou et al. / Agricultural Water Management 140 (2014) 37–47 39

ental s

vwsPaUf

ip

Fig. 1. Location of the experim

The theory and practice of SF gauging have been described pre-iously by Smith and Allen (1996). After installation, the probesere sealed with silicon foam to prevent rain water intrusion and

hielded with aluminum foil to avoid external thermal influences.robe signals were read at 10 s intervals and recorded as 10 minverages on a CR1000 data logger (Campbell Scientific, Logan, UT,SA). The data were downloaded weekly from Jul 15 to Oct 1, 2011,

or analysis.This study used the stem heat balance method (SHB) described

n detail by Kigalu (2007) to measure SF or transpiration rate in thelant stem. The SHB method uses a small heater wrapped around

180cm

900c

SWC, Trime tube

38 23'33.87''N109 11'58.27'' EElevation:1282 m

Flow sensor

SFM, Flow sensor Maize stem

Flow sensor Flow sensor

Fig. 2. Layout of the experiment

ite in Shaanxi Province, China.

the plant stem to heat the section. If a known amount of heat issupplied to a stem section and a constant amount of sap (water)flows through the section during transpiration, the temperature ofthe sapwood will reach a steady value that is inversely proportionalto the water flow. Energy balance was expressed as

Qf = Pin − Qcd − Qr + Qs (2)

where Qf is the amount of heat (W) transported in the moving sap;Pin is the heater power input (W); Qcd is the heat conduction lossalong the stem up- and downstream along the plant stem (W); Qr

m

Evaporation, Micro-lysimeter

Flow sensor

Trime tube Trime tube

Flow sensor

al setup in the maize field.

40 L. Hou et al. / Agricultural Water Management 140 (2014) 37–47

leaf ar

it

a

F

w(t

f(l(niattlomism

a

Fig. 3. Changes in (a) crop height and (b)

s the radial heat conduction loss (W); and Qs is the heat stored inhe stem section (W).

The SF velocity F (cm3 h−1) was calculated from the energy bal-nce across the sap flow meter as:

= Pin − Qcd − Qr + Qs

Cs × dTsap(3)

here Cs is the specific heat capacity of the sap or water4.186 J g−1 C−1) and dTsap is the temperature differential betweenhe heater and the stem section (◦C).

Measurements of Ep were taken using micro-lysimeters adaptedrom Boast and Robertson (1982) as described by Flumignan et al.2011) after testing them for accuracy and reliability. The micro-ysimeters comprised a double structure of rigid organic glass ringsinner ring: 0.102 m internal diameter; outer ring: 0.123 m inter-al diameter; both rings: 0.18 m long, 0.005 m wall thickness). The

nternal ring was filled by driving it through the soil, extracting it,nd removing excess soil before sealing the bottom with a thin plas-ic cap to prevent drainage. Then, the external ring was installed inhe same hole, the exposed soil surface was covered with a thinayer of rice husk, and the inner ring was installed into the outerne. The double organic glass structure was used to both facilitateicro-lysimeter extraction for daily weighing (Allen, 1990) and to

nsulate against thermal conductivity between the outer and inner

oil (Evett et al., 1995). Soil evaporation was measured using sixicro-lysimeters randomly arranged in the experimental plot.Therefore, in the absence of direct measurements of ETc and

ssuming negligible daily heat storage in small diameter plant

ea index (LAI) of maize in 2011 and 2012.

sections (Cammalleri et al., 2013), the ETc under an absence ofwater deficit was estimated as the sum of Tp and Ep measurements:

ETc = Ep + Tp (4)

In which ETc is crop evapotranspiration under an absence ofwater deficit, Ep is potential soil evaporation, and Tp is potentialtranspiration by SF measurement.

2.5. Weather data and soil water measurements

Meteorological variables, including air temperature, humidity,barometric pressure, radiation, rainfall, and wind speed/direction,were measured with an automatic meteorological station locatedat the experimental site, similar to a setup used by Cammalleriet al. (2013). Wind speed (m s−1) was measured with a windmonitor (05130-5; R.M. Young Co., Traverse City, MI, USA) 3.25 mabove the ground. Air temperature (◦C) and relative humidity, RH(%) were measured with a thermo-hygrometer (HMP45C; Vaisala,Helsinki, Finland) 3.12 m above the ground. Net radiation (W m−2)was measured using a NR-LITE sensor (Kipp & Zonen, Delft, TheNetherlands). Precipitation (mm h−1) was recorded using a raingauge (type 52203, R.M. Young Co.). All parameters were sampledevery 60 s and recorded as 1 h averages with an automatic datalogger (CR3000; Campbell Scientific).

Soil water content was measured in three locations every 5 daysusing three equally-spaced Trime TDR-IPH probes (IMKO Micro-modultechnik GmbH, Ettlingen, Germany) at depths of 10, 20, 30,40, 50, 60, 70, 80, and 90 cm below the soil surface.

ter Ma

2c

(

E

wn(s((

a

E

th2adabpFatflw

riua

K

K

itf

tt

K

2

s

G

L. Hou et al. / Agricultural Wa

.6. Calculation of reference crop evapotranspiration and cropoefficient

ET0, was calculated by the FAO Penman–Monteith equationAllen et al., 1998):

T0 =0.408� (Rn − G) +

(900�/ (T + 273)

)u2 (es − ea)

� + �(1 + 0.34u2)(5)

here ET0 is the reference evapotranspiration (mm d−1), Rn is theet radiation at the crop surface (MJ m−2 d−1), G is the soil heat fluxMJ m−2 d−1), T is the average air temperature (◦C), u2 is the windpeed at a height of 2 m (m s−1), es − ea is the vapor pressure deficitkPa), � is the slope of the vapor pressure and temperature curvekPa ◦C−1), � is the psychrometric constant (kPa ◦C−1).

ETc under standard conditions can be calculated from Kc and ET0s equation (6) (Allen et al., 1998):

Tc = Kc × ET0 (6)

ETc under standard conditions is usually determined by theime-averaged Kc approach. “Standard conditions” refer to a semi-umid climate with a relative humidity of 45%, wind speed of

m s−1, sufficient water, appropriate management, normal growth,nd a large area of high-yield crops. In FAO-56, the Kc curve isivided into four growth stages: initial, development, midseason,nd late season. The initial and midseason periods are characterizedy horizontal line segments and the development and late seasoneriods are characterized by rising and falling line segments. TheAO-56 Kc recommendations for maize during the midseason andt the time of harvest are Kc mid(Tab) = 1.2 and Kc end(Tab) = 0.6, respec-ively (Allen et al., 1998). In the study area, the initial period wasrom Apr 30 to May 27, the midseason from Jul 2 to Aug 20, and theate stage from Aug 21 to Oct 1 in 2011. The corresponding stages

ere May 3–31, Jul 1–Aug 19, and Aug 20–Sep 27 in 2012.The FAO-56 crop coefficients must be amended for different

egions according to the moisture and climatic conditions. With fullrrigation and in climates where RHmin differs from 45% or where2 is larger or smaller than 2.0 m s−1, the values Kc mid and Kc end,dapted to the actual climate conditions can be evaluated as:

c mid = Kc mid(Tab) + [0.04 (u2 − 2) − 0.004 (RHmin − 45)](

h

3

)0.3

(7)

c end = Kc end(Tab) + [0.04 (u2 − 2) − 0.004 (RHmin − 45)](

h

3

)0.3

(8)

where RHmin is the mean daily minimum relative humidity dur-ng the mid-season growth stage (%) for 20% < RHmin < 80%; and h ishe average plant height during the mid- or end-season stage (m)or 0.1 m < h < 10 m.

On the other hand, measured ETc fluxes were obtained ashe sum of plant transpiration and soil evaporation determinedhrough equation (4) in 2011. The sensor-based values of Kc were:

c = ETc

ET0(9)

.7. Groundwater contribution to maize water use

The groundwater contribution to maize water use or drainageurplus was computed using the water balance equation:

= ETc − P − I − �S (13)

nagement 140 (2014) 37–47 41

where G is groundwater contribution or drainage surplus (mm), P isprecipitation (mm), I is irrigation depth (mm), and �S is soil waterstorage change (mm).

2.8. Multiple linear regression

Transpiration of maize (Tp) in relation to climatic factors may beexpressed as follows:

Tp = b0 + b1 × Rn + b2 × T + b3 × RH + b4 × u2 (14)

where Tp is predicted transpiration of maize (cm3 h−1), b0, b1, b2,b3 and b4 are the regression coefficients of the multiple regressionequation.

3. Results and discussion

3.1. Variation of the reference evapotranspiration

Variation in maize stem Tp involves both internal and externalfactors. Internal factors include canopy structure, stomata opening,stem and root hydraulic conductivity, and hydraulic architec-ture characteristics; external factors are environmental, includingmeteorology and soil moisture (Cammalleri et al., 2013). Fig. 4ashows the dynamics of ET0, P, and I in 2012. Fig. 4b shows thesoil water content in the root zone (0–90 cm) in the same sea-son. The water content of each soil layer was higher than theSWCwp of 0.045 m3 m−3. The total available soil water in mm(TAW = 1000 × (SWCfc − SWCwp) × Zr) (Cammalleri et al., 2013) was230 mm. According to Table 22 of the FAO manual (Allen et al.,1998), p can be assumed to be equal to 0.50 under the maximumroot depth of 0.8–1.2 m for maize. The readily available soil waterin the root zone (RAW) was 115 mm evaluated as a percentage “p”of the TAW (RAW = p × TAW). In addition, the root zone depletion(Dr) was calculated according to the daily soil water balance. Thedepletions in the root zone and in the top soil surface, during thetwo growing seasons, were always lower than RAW (Dr < RAW).The absence of water deficit was verified during the two grow-ing seasons, that was Ks = 1. The RAW was suitable for maize crops,under the optimum soil water conditions (Allen et al., 1998). Con-sequently, it was possible to assume that the terms ETc which wasmeasured as the sum between plant transpiration (Tp) and soilevaporation (Ep), really represents the potential crop evapotranspi-ration (ETp). That is ETc = ETp = ETa. Then Kc can be calculated as theratio between ETc and ET0 (evaluated with the Penman–Monteithequation).

3.2. Diurnal variation in sap flow and climate variables

Transpiration rate (mm h−1) of the maize was estimated bydividing the average SF measured in six stems by the plant area(Cammalleri et al., 2013). In this study, the distance betweenthe maize stems was 37.5 cm, and the distance between rowswas 42.5 cm (Fig. 2). Therefore, each maize plant had an area of∼37.5 × 42.5 cm2. Differences in SF may be partially attributableto the measurement process. One way of reducing the measure-ment error of SF is to use several sensors in a plot instead of oneand averaging the measurements (Liu et al., 2008). The mean dailytranspiration rates based on sensor readings were 6.6, 5.7, 2.9, 1.7,and 1.2 mm d−1 for Jul 15–16, Aug 2–3, Aug 26–27, Sep 24–25, andSep 27–28, respectively.

The Tp of maize had an obvious day–night rhythm synchronouswith solar radiation. Fig. 5 shows the average Tp by SF measurement

for two consecutive days during five different growing periods.The Tp shows a clear diurnal pattern, starting at 8:00 AM, peakingaround noon, and stopping around 8:00 PM. Tp also varied through-out the growing season; it was very high in mid-July and early

42 L. Hou et al. / Agricultural Water Management 140 (2014) 37–47

F transpf

ASLea

icmo(tltthhf

iarto

T

Fig. 7 illustrates the measured Tp, Ep and ETc, as well as P andI from Jul 15 to Oct 1, 2011. The maximum daily ETc (9.0 mm d−1)

ig. 4. Temporal dynamics of (a) precipitation, irrigation, and FAO reference evaporom May 23 to Sep 27, 2012.

ugust, but decreased by half in late August and stopped in lateeptember, when the maize was ripe for harvest. In Jul 15–16, 2011,AI and crop height were almost constant (Fig. 3), and the differ-nces observed in sap fluxes were due to meteorological variabilitynd soil water content.

Climatic conditions were the predominant factors influenc-ng Tp. Fig. 6 shows the hourly Tp measurements in relation tolimatic factors. The diurnal fluctuation in Tp was mainly deter-ined by net radiation, as indicated by the corresponding peaks

f these two measurements and their high correlation coefficient>0.95; Table 1) in all growing periods. These data also supporthe results of (Liu et al., 2008), in the sense that the Tp wasinearly related to solar radiation. Tp was also correlated to airemperature, and peak Tp did not occur at the highest tempera-ure (from 14:00–16:00). Tp was negatively correlated with relativeumidity, but peak Tp did not occur at the minimum relativeumidity. Wind speed has less influence on Tp than the other

actors.Crop water requirements are partly determined by meteorolog-

cal factors, mainly solar radiation, air temperature, air humidity,nd wind speed. For example, on Aug 2–3, 2011, the multiple linearegression between Tp of these four factors accounted for >97% ofhe variation in measured maize Tp. The multiple linear regression

n Aug 2–3, 2011, was:

p = −12.0559 + 0.1095Rn − 7.0893T − 0.036RH

+1.7767u2, R2 = 0.9762 (15)

iration (ET0) and (b) measured soil water content from May 30 to Oct 1, 2011, and

3.3. Measured transpiration and soil evaporation

Fig. 5. Average transpiration (Tp) of maize by sap flow (SF) measurement throughthe day during different growing periods in 2011. Vertical bars indicate standarderror.

L. Hou et al. / Agricultural Water Management 140 (2014) 37–47 43

Fig. 6. Influence of climate factors on transpiration (T) of maize by sap flow (SF)measurement on Aug 2–3, 2011.

Fig. 7. Temporal variation in soil evaporation, crop transpiration, and total evapo-transpiration during the study period in 2011. Precipitation and irrigation are alsoshown.

and Tp (7.2 mm d−1) occurred on Jul 15. Soil evaporation accountedfor only 20% of total ETc, therefore, variation in ETc was mainlydetermined by changes in Tp during this high-LAI stage (Fig. 3).After late August, ETc declined gradually, mainly because Tp

decreased in the late growing period (Fig. 7). Total ETc from Jul 15to Oct 1 in 2011 was 330 mm; 245 mm (74%) was due to Tp and85 mm (26%) was due to Ep. The mean daily ETc was 4.17 mm d−1

during the experimental period.

3.4. Maize crop coefficient

The single crop coefficient during the initial stage (Kc ini) wasestimated based on the equation and figure of FAO-56 (Allen et al.,1998). The proposed Kc mid and Kc end for maize was estimated usingthe meteorological data in the study area in 2011, as shown in Eqs.(7) and (8) and in Fig. 8 (symbols are observed Kcs, line series aresimulated Kcs). They were 0.34, 1.23, and 0.59, respectively. TheFlow32-1K Sap Flow System gauge probes were only installed whenthe stem diameter was large enough to meet the manufacturer rec-ommendations, so SF data were not available for the initial growthperiod. The crop coefficients using Tp (measured SF) and soil evap-oration data only for middle and late growth stages, are shown by

the solid dots in Fig. 8. The measured daily evapotranspiration werethe true values. The estimated ETc values follow a similar trendof the corresponding measured values, and the two series fitted

Fig. 8. Comparison between the single maize crop coefficients determined in theexperiment and the FAO-56 recommendations.

44 L. Hou et al. / Agricultural Water Management 140 (2014) 37–47

Table 1Correlation coefficients between maize sap flow and climatic factors in 2011.

Correlation coefficient Net radiation (W m−2) Temperature (◦C) Relative humidity (%) Wind speed (m s−1)

Jul 15–16 0.93 0.60 −0.58 −0.09Aug 2–3 0.97 0.61 −0.64 0.57Aug 26–27 0.97 0.59 −0.61 0.52Sep 24–25 0.98 0.67 −0.63 0.67Sep 27–28 0.90 0.75 −0.63 0.67Whole period 0.95 0.64 −0.618 0.47

* aize saa ite.

wiKTtaaoc2a

Ed

Fe

Pereira (2000) showed that the Kc values calculated by the FAOmethods were close to those estimated from the soil water bal-

Stem sap flow (cm3 h−1). n = 48 h × 5(4 climate factors plus sap flow data) = 240. Mttained through an automatic meteorological station located at the experimental s

ell. So, the maize coefficient Kc through Eqs. (7) and (8) was val-dated. The estimated Kc values for maize in the 2012 season werec ini = 0.39, Kc mid = 1.13, and Kc end = 0.54 through Eqs. (7) and (8).he differences observed in 2011 and 2012 on Kc values obtained athe different phenological stages was because the Kc for mid-seasonnd late season maize increased as relative humidity increased ands wind speed decreased in 2012. More arid climates and conditionsf greater wind speed had higher Kc values in 2011. More humidlimates and conditions of lower wind speed had lower Kc values in012. The Kc ini was 0.37, Kc mid was 1.18, and Kc end was 0.57 whenveraging the Kc values for the two experimental seasons.

The comparison between the estimated and measured maizeTc is presented in Fig. 9. The coefficient of determination betweenaily maize ETc estimated using the FAO-56 Kc values and that

ig. 9. Comparison of (a) daily and (b) cumulative estimated and measured maizevapotranspiration (ETc) values in 2011.

p flow in six plants was measured using Flow 32 sensors. The climatic factors were

determined using the measurements of this study was 0.85. And,the estimated ETc was 336.78 mm in 79 days (4.26 mm d−1 onaverage), was similar to the Flow 32 and micro-lysimeter ETc at329.66 mm in 79 days (4.17 mm d−1 on average). So, the calculatedETc agreed well with the measured values. The measured daily Tp

by SF measurement, and micro-lysimeter results can be consideredaccurate estimates of daily evapotranspiration in the maize field.Therefore, the applicability of the calculated maize crop coefficientsduring the growing season in this area was confirmed. Liu and

ance calculation data in the North China Plain, and they concludedthat the single crop-coefficient method can be used to determine

Fig. 10. Daily evapotranspiration during the 2011 (Apr 30 to Oct 1) and 2012 (May 3to Sep 27) growing seasons, calculated according to the Penman–Monteith equationand a crop coefficient (Kc) calibrated using the FAO approach and using actual sapflow and soil evaporation data.

L. Hou et al. / Agricultural Water Management 140 (2014) 37–47 45

F and (b

tCstscmT1fiaaa

3p

tEai

ig. 11. Maize (a) cumulative crop evapotranspiration (ETc) and water table depth

he crop coefficients for winter wheat and summer corn in Northhina. On the other hand, Tolk and Howell (2000) found that theingle crop coefficient (Kc) procedure underestimated final cumula-ive daily measured evapotranspiration (ETc) in fully-irrigated grainorghum treatments by an average of 9%. However, the single crop-oefficient method using Kc, which is the ratio of crop ETc to ET0, isuch simpler and more convenient and is used widely in practice.

o obtain ETc in the Shaanxi province, Northwest China (34◦20′N,08◦24′E), for fields with a loess loam soil via the single crop coef-cient method, Kang et al. (2003) measured ETc using lysimetersnd ET0 using the Penman–Monteith method. They then estimatedverage, maximum, and minimum seasonal Kc values of 1.04, 1.43,nd 0.45, respectively, for maize.

.5. Crop evapotranspiration changes throughout the growtheriod

Using the averaged Kc values from this study, maize ETc during

he 2011 and 2012 growing seasons was calculated. Daily values ofTc from Apr 30 to Oct 1 in 2011 and from May 3 to Sep 27 in 2012re presented in Fig. 10. The total ETc values were 583 and 500 mmn the two growing seasons, respectively. These values were similar

) soil water content during a relative dry period from Aug 8 to Aug 15, 2011.

to the water use of maize (581 mm) measured in a similar regionin the Maowusu Sandlot by Guo et al. (2005).

Seasonal variation of ETc indicated that the evapotranspirationrate was very low in the early growing stage (Apr 30 to the end ofMay) when LAI was small (Fig. 3). ETc increased gradually in June asLAI became greater. In late July to August, ETc generally peaked, asdid plant height and LAI. ETc dropped in September when the plantsneared harvest. During days with rainfall, ETc was almost zero.

3.6. Groundwater contribution to water use of maize

Water for plant transpiration comes from soil water, whichis directly linked to rainfall and irrigation, and potentially fromgroundwater (Xu et al., 2010). Fig. 11 shows the decrease in soilwater content (especially from 80 to 90 cm depth) and the decreasein water table depth caused by ETc during the dry period from Aug8–15, 2011.

Crop-water use of shallow groundwater depends on several fac-

tors, including groundwater table depth, soil hydraulic properties,crop growth stage, weather, and irrigation. The total ETc rate ofmaize during the maize growing season of 155 days was ∼583 mmin 2011, with 214 mm of precipitation, 177 mm of irrigation, and

4 ter Ma

1tfmwsgtos(wd(ltidP

batct

4

cmNtc

aa(38

mtaaopm0vrrw

Tctsec

A

e

6 L. Hou et al. / Agricultural Wa

92 mm of groundwater of that same period. Groundwater con-ributed to 33% of ETc with an average 1.12 m groundwater table andull irrigation in 2011. In 2012, the total evapotranspiration rate of

aize during the maize growing season of 148 days was ∼500 mm,ith 411 mm of precipitation, 60 mm of irrigation, an increase in

oil water storage of 146 mm, and 175 mm of groundwater. Thus,roundwater contributed to 27% of ETc at the average groundwa-er table of 0.89 m with full irrigation in 2012. The capillary riser groundwater contribution to ETc were considered negligible intudies by Liu and Pereira (2000) in Hebei Province, North Chinasilt loam soil) and by Kang et al. (2003) in Shaanxi Province, North-est China (loess loam soil), where the groundwater table waseeper and the soil had a higher clay content. Luo and Sophocleous2010) assessed the seasonal groundwater contribution to ETc usingysimeter observations and numerical experiments, and they foundhat the extinction depth of groundwater evaporation was ∼3.8 mn a multilayered soil profile grouped into six soils, with slightifferences among the irrigation scenarios in Yucheng, Shandongrovince, China.

Therefore, the higher groundwater contribution to ETc occurredecause the average depth of the groundwater table in the researchrea was only 1.12 m in 2011 and 0.89 m in 2012, and the soil con-ained relatively little clay and more sand. Therefore, groundwateran be a significant source of water for crops where the groundwa-er table is shallow, as in the Hailiutu River catchment.

. Conclusions

The main objective of this study was to assess the utility of therop coefficients for maize suggested by FAO using actual measure-ents for a typical semi-arid area of the Hailiutu River catchment inorthwest China under conventional surface irrigation, as well as

heir relationships to meteorological conditions and soil moistureontent.

An experimental dataset of Tp and Ep measurements under anbsence of water deficit, collected from Jul 15 to Oct 1 2011, wasnalyzed. The plant transpiration measured by sap flow sensorsby the stem heat balance method) was 245 mm (74% of total ETc of30 mm) and soil evaporation measured by micro-lysimeters was5 mm (26% of total ETc) from Jul 15 to Oct 1 in 2011 growing season.

The research showed that joint measurements of sap flow,icro-lysimeter techniques, and the FAO-56 approach allows a bet-

er understanding of the combined roles of actual Tp and Ep. In thebsence of direct measurements of Tp before Jul 15, the FAO-56pproach calculation, coupled with simultaneous measurementsf soil water content, ensured the correct value for evapotrans-iration fluxes. Based on the transpiration and soil evaporationeasurements, the single crop coefficients for this region were

.37, 1.18, and 0.57 during the initial period, midseason, and at har-est, respectively. The estimated and measured evapotranspirationates agreed well. Using these crop coefficients, total evapotranspi-ation in the maize field during the 2011 and 2012 growth seasonsas estimated to be 583 and 500 mm, respectively.

Groundwater contributed 27–33% of maize ETc in this study.herefore, groundwater is a significant source of water to meetrop demands in this area of the Hailiutu River catchment, wherehe groundwater table is shallow and the soil contains amounts ofand. Further experiments will be conducted in the coming years toxtend the experimental dataset to periods with different soil waterontents to verify the crop coefficients under multiple conditions.

cknowledgments

The research was partly supported by the National Natural Sci-nce Foundation of China (No. 41172225) and the Fundamental

nagement 140 (2014) 37–47

Research Funds for the Central Universities of China (2652012068).We also thank the Dutch government for funding the project “AsiaFacility for China, Partnership for Education and Research in Waterand Ecosystem Interactions”. The authors acknowledge the twoanonymous reviewers and the chief editor for their constructivecomments, which improved the overall quality of the manuscript.

References

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration: guide-lines for computing crop water requirements. In: FAO Irrigation and DrainagePaper No. 56, Rome, Italy, p. 300.

Allen, S.J., 1990. Measurement and estimation of evaporation from soil under sparsebarley crops in northern Syria. Agric. For. Meteorol. 49, 291–309.

Boast, C.W., Robertson, T.M., 1982. A micro-lysimeter method for determining evap-oration from bare soil: description and laboratory evaluation. Soil Sci. Soc. Am.J. 46, 689–696.

Cammalleri, C., Agnese, C., Ciraolo, G., Minacapilli, M., Provenzano, G., Rallo, G., 2010.Actual evapotranspiration assessment by means of a coupled energy/hydrologicbalance model: validation over an olive grove by means of scintillometry andmeasurements of soil water contents. J. Hydrol. 392, 70–82.

Cammalleri, C., Rallo, G., Agnese, C., Ciraolo, G., Minacapilli, M., Provenzano, G., 2013.Combined use of eddy covariance and sap flow techniques for partition of ETfluxes and water stress assessment in an irrigated olive orchard. Agric. WaterManage. 120, 89–97.

Chabot, R., Bouarfa, S., Zimmer, D., Chaumont, C., Moreau, S., 2005. Evaluation of thesap flow determined with a heat balance method to measure the transpirationof a sugarcane canopy. Agric. Water Manage. 75, 10–24.

De Bruin, H.A.R., 2002. Introduction: Renaissance of scintillometry. Bound.-Lay.Meteorol. 105, 1–4.

Droogers, P., 2000. Estimating actual evapotranspiration using a detailed agro-hydrological model. J. Hydrol. 229, 50–58.

Du, Z.G., 2011. Maize variety yield experiment in sand area of Yulin District. AnhuiAgric. Sci. Bull. 17 (15), 89–111 (in Chinese).

Evett, S.R., Warrick, A.W., Mathias, A.D., 1995. Wall material and capping effects onmicrolysimeter temperatures and evaporation. Soil Sci. Soc. Am. J. 59, 329–336.

Feddes, R.A., Kowalik, P.J., Malinka, K.K., Zaradny, H., 1978. Simulation of Field WaterUse and Crop Yield. John Wiley and Sons, PUDOC, Wageningen, pp. 189–199.

Flumignan, D.L., Faria, R.T., Prete, C.E.C., 2011. Evapotranspiration components anddual crop coefficients of coffee trees during crop production. Agric. Water Man-age. 98, 791–800.

Green, S.R., Kirkham, M.B., Clothier, B.E., 2006. Root uptake and transpiration: frommeasurements and models to sustainable irrigation. Agric. Water Manage. 86,165–176.

Guo, K.Z., Li, H.P., Shi, H.B., Yang, Y.S., Xu, B., Tong, C.F., 2005. Study on water rate offodder grass and optimization of saving water irrigation program in Maowususandlot. J. Irrig. Drain. 24 (1), 24–27 (in Chinese with English abstract).

Homaee, M., Feddes, R.A., Dirksen, C., 2002. Simulation of root water uptake. I.Non-uniform transient salinity using different macroscopic reduction functions.Agric. Water Manage. 57, 89–109.

Kang, S.Z., Gu, B.J., Du, T.S., Zhang, J.H., 2003. Crop coefficient and ratio of transpira-tion to evapotranspiration of winter wheat and maize in a semi-humid region.Agric. Water Manage. 59 (3), 239–254.

Kigalu, J.M., 2007. Effects of planting density on the productivity and water use of tea(Camellia sinensis L.) clones. I. Measurement of water use in young tea using sapflow meters with a stem heat balance method. Agric. For. Meteorol. 90, 224–232.

Li, K.Y., De Jong, R., Coe, M.T., Ramankutty, N., 2006. Root-water-uptake based upona new water stress reduction and an asymptotic root distribution function. EarthInteract. 10, 1–22.

Liu, H.J., Cohen, S., Tanny, J., Lemcoff, J.H., Huang, G.H., 2008. Transpiration estimationof banana (Musa sp.) plants with the thermal dissipation method. Plant Soil 308,227–238.

Liu, H.J., Yu, L.P., Luo, Y., Wang, X.P., Huang, G.H., 2011. Responses of winter wheatevapotranspiration and yield to sprinkler irrigation regimes. Agric. Water Man-age. 98, 483–492.

Liu, Y., Pereira, L.S., 2000. Validation of FAO methods for estimating crop coefficients.Trans. CSAE 16 (5), 26–30 (in Chinese with English abstract).

Luo, Y., Sophocleous, M., 2010. Seasonal groundwater contribution to crop-wateruse assessed with lysimeter observations and model simulations. J. Hydrol. 389,325–335.

Perrier, A., Tuzet, A., 1991. Land surface processes: description, theoreticalapproaches and physical laws underlying their measurements. In: Schmugge,T.J., André, J.C. (Eds.), Land Surface Evaporation—Measurement and Parameter-ization. Springer-Verlag, New York, NY, pp. 145–155.

Rallo, G., Provenzano, G., 2012. Modelling eco-physiological response of table olivetrees (Olea europaea L.) to soil water deficit conditions. Agric. Water Manage.120, 79–88.

Robertson, J.A., Gazis, C.A., 2006. An oxygen isotope study of seasonal trends in soilwater fluxes at two sites along a climate gradient in Washington State (USA). J.Hydrol. 328, 375–387.

Swanson, R.H., 1994. Significant historical developments in thermal methods formeasuring sap flow in trees. Agric. For. Meteorol. 72, 113–132.

ter Ma

S

S

T

W

ponents: sap-flow, soil water budget, eddy covariance and catchment water

L. Hou et al. / Agricultural Wa

mith, D.M., Allen, S.J., 1996. Measurement of sap flow in plant stems. J. Exp. Bot. 47(305), 1833–1844.

tockle, C.O., Jara, J., 1998. Modeling transpiration and soil water content from a corn(Zea Maize L.) field: 20 min vs. daytime integration step. Agric. For. Meteorol. 92,119–130.

olk, J.A., Howell, T.A., 2000. Measured and simulated evapotranspiration of grainsorghum grown with full and limited irrigation in three high plains soils. Trans.ASAE 44 (6), 1553–1558.

enninger, J., Beza, D.T., Uhlenbrook, S., 2010. Experimental investigations ofwater fluxes within the soil–vegetation–atmosphere system: stable isotope

nagement 140 (2014) 37–47 47

mass-balance approach to partition evaporation and transpiration. Phys. Chem.Earth 35 (13–14), 565–570.

Wilson, K.B., Hanson, P.J., Mulholland, P.J., Baldocchi, D.D., Wullschleger, S.D., 2001. Acomparison of methods for determining forest evapotranspiration and its com-

balance. Agric. For. Meteorol. 106, 153–168.Xu, X., Huang, G., Qu, Z., Pereira, L.S., 2010. Assessing the groundwater dynamics and

predicting impacts of water-saving in the Hetao Irrigation District Yellow Riverbasin. Agric. Water Manage. 98, 301–313.