Agricultural Water Management 182 (2017) 139150

Contents lists available at ScienceDirect

Agricultural Water Management

journal homepage: www.elsevier.com/locate/agwat

Root development and water uptake in winter wheat under different

irrigation methods and scheduling for North China
Shiva K. Jha a,b , Yang Gao a , Hao Liu a , Zhongdong Huang a , Guangshuai Wang a,b ,
Yueping Liang a,b , Aiwang Duan a,
a
Key Laboratory of Crop Water Use and Regulation, Ministry of Agriculture Farmland Irrigation Research Institute, Chinese Academy of Agricultural
Sciences, Xinxiang, Henan 453003, PR China
b
Graduate School of Chinese Academy of Agricultural Sciences (GSCAAS), Beijing 100081, PR China

a r t i c l e i n f o a b s t r a c t

Article history: A eld experiment was conducted on winter wheat (Triticum aestivum L.) during 20132014 and
Received 1 January 2016 20142015 to study the root distribution prole and soil water dynamics under the main currently
Received in revised form used irrigation methods in the North China Plain (NCP). The WinRHIZO system and the HYDRUS-1D
14 December 2016
model were used to identify a promising irrigation schedule. In this two-factor experiment, three irri-
Accepted 16 December 2016
gation methods, i.e., sprinkler irrigation (SI), surface drip irrigation (SDI) and surface ooding (SF), were
scheduled to irrigate the crop as soon as the soil water content decreased to 70%, 60% and 50% of the
Keywords:
eld capacity. The results showed that both the irrigation method and irrigation schedule inuenced
Soil water dynamics
Root morphology
root development, the prole root distribution pattern and the prole root water uptake (RWU). The soil
Root growth surface temperature uctuated very rapidly depending on the irrigation method and scheduling system
Irrigation management used, whereas prole soil temperature uctuations became more consistent with depth. The RWU was
Soil temperature higher in the upper soil layer (060 cm) for all irrigation methods for frequently irrigated treatments, and
the maximum was observed in SDI compared to SI and SF due to the higher root length density (RLD) in
the top soil under SDI. On the other hand, the RWU was higher in SF at a deep soil prole below 60 cm,
where it had a higher RLD compared to that of SI and SDI. SDI at 60% of FC not only improved water
uptake but also resulted in better water productivity and produced the highest grain yield (9.53 t/ha).
The simulated RWU and soil water dynamics presented in this paper will be helpful to improve winter
wheat production in the NCP and can be used as a reference for further research on water management
practices.
2016 Published by Elsevier B.V.

1. Introduction
Sustainable management of the regional groundwater in North
China, where more than 70% of fresh water is used in agriculture, is
a major challenge (Du et al., 2015). This is why severe groundwater
level decreases (0.53 m/year) were found by Currell et al. (2012)
throughout the North China plain (NCP) in the last three to four
decades. Liu et al. (2002) and Sun et al. (2006) stated that winter
precipitation ranges from 50 mm in dry to 150 mm in wet years
in areas where the water requirement for winter wheat produc-
tion is estimated to be 300450 mm. Thus, maximizing the water
Corresponding author at: Farmland Irrigation Research Institute, Chinese
Academy o Agricultural Sciences, No. 380 Hongli Road, Xinxiang, Henan 453003,
PR China.
E-mail address: duanaiwang@aliyun.com (A. Duan).
productivity of winter wheat is the current basic need for NCP
farmers. This can be achieved by proper soil water management
practices. Selecting water management practices without knowing
the soil water dynamics and water uptake by the crop makes it
nearly impossible to support an efcient winter wheat production
system that includes root development (Ritchie, 1981; Samson and
Sinclair, 1994). Less attention has been paid to studies of root mor-
phology because roots occur beneath the soil surface and require
tedious measurements (Ephrath et al., 1999; McMichael and Taylor,
1987). Roots are anchors and function as entry points for the uptake
of water and mineral nutrients for crops. Roots also serve as sen-
sors of water stress and have many synthetic functions of shoot
cells (Kramer and Boyer, 1995). Thus, a better understanding of root
morphological growth and root water uptake (RWU) patterns in the
soil prole is very important for successful crop growth and max-
imum grain production (Coelho and Or, 1999; Roose and Fowler,
2004; Samson and Sinclair, 1994). Zuo et al. (2004) collected wheat

http://dx.doi.org/10.1016/j.agwat.2016.12.015
0378-3774/ 2016 Published by Elsevier B.V.
140 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
root length density (RLD) information from several studies and con-
cluded that RLD is an important parameter to model water and
nutrient movement in the vadose zone as well as to study soil-
root-shoot-atmosphere interactions. Carvalho and Foulkes (2013)
also concluded that the RLD (cm/cm3 ) is the most suitable param-
eter to describe water uptake by plant roots compared with other
root traits. Thin roots have a relatively high specic root length
(SRL) or length: dry weight ratio, whereas the ne root system
functions as the principal pathway for water and nutrient absorp-
tion (Eissenstat, 1992). This author further concluded that water
and nutrient uptake increases are more likely to occur with greater
root length than root mass and provided evidence that a higher SRL
tended to lead to greater plasticity in root growth and a greater
physiological capacity but shorter root longevity and less myc-
orrhizal dependency. Various techniques have been developed to
monitor root dynamics under eld conditions where excavation
techniques, including soil cores, have long been considered to pro-
vide the most reliable estimates of root morphology (Samson and
Sinclair, 1994).
Many researchers have correlated root growth and water uptake
under different soil moisture regimes (Carvalho and Foulkes, 2013;
Coelho and Or, 1999). The amount of soil water absorbed by plant
roots not only depends on the physical characteristics of the soil but
also has a greater inuence on prole root growth features (Yang
et al., 2006). Non-irrigated plants have thicker roots, fewer roots
near the soil surface, and more roots at deeper depths (Rowse,
1974). The RWU is distributed over the root zone according to
the spatial root distribution and is controlled by climatic demand,
the spatial distribution of soil water availability and root density
unek
(Albasha et al., 2015; Sim and Hopmans, 2009). Li et al. (2014)
concluded that the RWU rate of maize increased after rainfall or irri-
gation. Xue et al. (2003) found that winter wheat roots continued
to grow up to a depth of 2 m until the booting stage and concluded
that the RWU rate decreased as available soil water decreased. The
root sampling results presented by Zhang et al. (2004) showed that
winter wheat has a prole root system with an average maximum
rooting depth of 2 m, and most of the root system is concentrated
in the upper 40 cm of soil, which is why the roots in the top layer
of soil play an important role in soil water uptake. Kmoch et al.
(1957) mentioned that the root development of winter wheat was
inuenced by soil moisture. These authors found that a dense net-
work of roots developed in the soil when the soil moisture tension
was above 15 atmospheres, and they observed roots at a depth of 13
feet under favorable moisture conditions. Maximum root growth in
the subsoil signicantly improved the soil water supply to the crop
by shifting root growth downward during the growing period due
to water depletion in the surface soil (Torreano and Morris, 1998).
On the other hand, soil temperature affects the growth, initiation,
branching, orientation, direction of growth, and root turnover of
root system components (Kaspar and Bland, 1992). These authors
showed that as the temperature increases, roots grow faster and
reach a maximum growth rate at approximately 30 C for maize and
pecans, above which the rate begins to decrease. The soil tempera-
ture prole distribution is greatly affected by the irrigation method
and undoubtedly inuences root water uptake directly or indirectly
(Lv et al., 2013a).
Coelho and Or (1999) emphasized that for irrigation schedul-
ing, it is necessary to consider the effect of the RWU rate on soil
water dynamics. Li et al. (2010) compared three irrigation sched-
ules applied during the jointing, heading and milking stages of
winter wheat and concluded that single irrigation event applied
during the jointing stage caused an increase in root length den-
sity in a >30 cm deep soil prole compared to 2 or 3 irrigation
events during other growth stages. Lv et al. (2010) concluded that
the irrigation method inuences winter wheat root development
and the water uptake prole, even for the same irrigation schedule.
These authors found that root uptake in the upper zone of the soil
prole increased with increases in irrigation frequency. An experi-
ment performed by Camposeo and Rubino (2003) on autumn sown
sugar beet clearly illustrated that the applied irrigation frequen-
cies signicantly affected RWU. These authors concluded that the
root length density along the soil prole decreased more than 76%
by decreasing the irrigation frequency. Above and below ground
biomass is more severely limited by water stress than by nutrient
stress (Fabio et al., 1995), whereas root initiation decreases with
increasing depth (Torreano and Morris, 1998) if adequate water
can be accessed by the surface roots. Although there are numer-
ous fractional studies on the effect of irrigation method (Lv et al.,
2010) or irrigation frequency (Li et al., 2010; Shao et al., 2009; Xue
et al., 2003) on root growth or water uptake, a combined approach
of irrigation method and scheduling system with water manage-
ment practices applied to an entire farmers eld to understand the
overall soil water dynamics and their inuencing parameters has
not been attempted.
Thus, the objectives of this eld-based winter wheat experiment
with a promoting irrigation method and feasible irrigation schedule
were to (i) measure the root morphological growth, (ii) study the
effects of soil moisture and temperature and (iii) estimate soil water
dynamics in the root zone under surface drip irrigation (SDI), sprin-
kler irrigation (SI) and surface ooding (SF), with irrigation at 50%,
60% and 70% of FC. The results of this study will provide references
for further research and the farmers of the NCP to design a practi-
cal and environmentally friendly irrigation system by selecting an
appropriate irrigation method with a proper irrigation schedule.
2. Materials and methods
2.1. Experimental site
Field research was conducted at the experimental station of
the Farmland Irrigation Research Institute (FIRI) of the Chinese
Academy of Agricultural Science (CAAS), located in Qiliying, Xinxi-
ang City of Henan Province in North China (35 08 N, 113 45 E and
81 m altitude). The experimental site has underground pipelines
connected to pumps with pressure regulators to supply ground
water as source of irrigation water at a desired pressure head and
a weather station is located in very close proximity to the experi-
mental plots. The seasonal precipitation from mid-October to early
June ranges from 60 to 200 mm, with seasonal water consumption
of approximately 450500 mm (Sun et al., 2006; Zhang et al., 2002)
and a mean seasonal air temperature that varies between 10 and
12 C.
2.2. Soil specication, tillage and harvesting
The soil characteristics, such as the physical and hydraulic
parameters, were investigated before the irrigation treatments
started and are presented in Table 1 for different root zone
depths. The average contents of the available soil nutrients, i.e.,
nitrogen (N), phosphorous (P), and potassium (K), at the experi-
mental site were 40.20, 11.90, and 100.51 mg kg1 , respectively.
The pH and electrical conductivity (EC) of the soil were 8.51 and
257.6 s cm1 , respectively, whereas the soil organic matter con-
tent was 1.64 g kg1 .
Seed beds were prepared by plowing to a depth of 20 cm using
a tractor drawn rotary cultivator; larger soil clods were smoothed
using a harrow to ensure a completely at bed. The equal amount
of basal fertilizer dose was N: 120 kg ha1 (50% of total N), applied
as ammonium nitrate, P: 90 kg ha1 , applied as calcium super-
phosphate, and K: 30 kg ha1 , applied as potassium sulfate in all
treatments (Gao et al., 2014). The remaining 50% of N (120 kg ha1 ),
 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150 141

Table 1
Soil specications (physical and hydraulic) of experiment eld.

Soil Depth Particle Size Distribution Texture B.D. FC r s (1/cm) n Ks Initial (cm3 /cm3 )
(cm) (g/cm3 ) (cm3 /cm3 ) (cm3 /cm3 ) (cm3 /cm3 ) (cm/day)

% Clay % Silt % Sand 12/3/2014 17/3/2015

020 3.80 43.14 53.06 Sandy Loam 1.56 0.3408 0.0302 0.4087 0.016 1.4629 55.85 0.132 0.143
2040 6.61 45.43 47.96 Loam 1.58 0.3076 0.0362 0.4033 0.011 1.5089 39.79 0.142 0.141
4060 6.06 48.33 45.61 Sandy Loam 1.54 0.3269 0.0358 0.4086 0.009 1.5402 43.98 0.204 0.194
6080 4.55 47.49 47.96 Sandy Loam 1.42 0.2831 0.0325 0.412 0.011 1.5161 52.57 0.246 0.271
80100 1.57 16.95 81.48 loamy Sand 1.45 0.2937 0.0340 0.3931 0.047 1.8359 131.2 0.235 0.300
Average 4.52 40.27 55.21 Sandy Loam 1.51 0.3104 0.0313 0.403 0.019 1.4435 51.81 0.1918 0.2098

Note: abbreviations B.D. stand for bulk density; FC for eld capacity; s and r for saturated and residual water content; is SWC at the rst day of simulation; and n for
coefcients in van Genuchten-Mualem equation; and Ks for saturated hydraulic conductivity.

was applied as ammonium nitrate by hand broadcasting in the
sprinkler irrigation and surface ooding treatments, immediately
before the rst irrigation event, whereas it was dissolved in water
and used as a top dressing during the rst irrigation event in the
drip irrigation treatment. The high-yielding winter wheat (Triticum
aestivum L.) cultivar Aikang 58 was sown at 180 kg ha1 (ensuring
350400 plants m2 at full germination), with 20 cm row spacing,
on the 20th of October 2013 and the 18th of October 2014, using a
tractor drawn seed-cum-fertilizer drill; the crops were harvested
on the 31st of May 2014 and the 6th of June 2015, respectively.
A harvest sample (1 m2 ) was randomly collected from each treat-
ment. Grain and straw were separated manually by rubbing and
cleaning the grains, which were sun dried to a 12% (wet basis)
moisture level (Gao et al., 2014; Jin et al., 2009).
2.3. Experimental design
Farmers in North China generally irrigate winter wheat in early
or late spring after winter dormancy depending on the soil moisture
conditions (Gao et al., 2014; Shao et al., 2011). Considering this
typical schedule, the experiments were designed to be conducted
from the turn-green stage (mid March) to harvest (end of May or
early June) for the 20132014 and 20142015 cropping seasons.
Two factorial treatments were arranged in a split-plot design.
The main plots included three irrigation schedules (i.e., irrigation
was applied as soon as the soil water content decreased to 50%, 60%
or 70% of eld capacity) considered as one factor, and the sub-plots
included three irrigation methods, i.e., sprinkler irrigation (SI), sur-
face drip irrigation (SDI), and surface ooding (SF). The treatment
designs were coded as follows:
Each main plot contained 3 sub-plots of 10 m 15 m separated
by a 1 m border on each side. The basic parameters of the sprin-
klers used in the sprinkler irrigation treatments included a working
pressure of 0.20.25 MPa, coefcient of uniformity (CU) of approx-
imately 90%, and sprinklers mounted on120 cm high risers to cover
a 5 m radius. In each sprinkler irrigated sub-plot, at each of the
four corners, sprinklers sprinkled at 90 (Q = 0.22 m3 /h), with 6
sprinklers sprinkling at 180 (Q = 0.3 m3 /h) at the outer edge and 2
sprinklers sprinkling at 360 (Q = 0.55 m3 /h) in the interior. Surface
drip laterals working at a pressure of 0.10.15 MPa were placed at
40 cm, i.e., between two alternate rows with emitters (Q = 2.2 L/h)
spaced 20 cm apart within each lateral. Surface ooding was carried
out using PVC pipes ( 2.5 ) with 2 nos. outlets ( 6 ; Q = 8 m3 /h)
facing upward and xed at 2/3rd and 1/3rd of the length of the
sub-plots designed for the surface ooding treatments.
2.4. Sampling and measurements
The daily weather data, such as the atmospheric temperature
(maximum and minimum), sunshine hours, wind speed, relative
humidity, and precipitation, were collected every day at 8:00
AM from the automatic weather station located in close proxim-
ity to the experimental area. The soil water content (SWC) was
determined weekly using TRIME (Time domain reectometry with
Intelligent Micro Elements, IMKO, Ettlingen, Germany) tube access
probes placed at with 20 cm deep intervals in the soil layer up
to a depth of 100 cm. The SWC of the top 020 cm soil layer was
determined using a gravimetric method. Some additional SWC
measurements were carried out before and after irrigation and dur-

S1 = Irrigate with a sprinkler as soon as soil moisture decreased to 50% of FC

S2 = Irrigate with a sprinkler as soon as soil moisture decreased to 60% of FC

S3 = Irrigate with a sprinkler as soon as soil moisture decreased to 70% of FC

D1 = Irrigate with a surface drip as soon as soil moisture decreased to 50% of FC

D2 = Irrigate with a surface drip as soon as soil moisture decreased to 60% of FC

D3 = Irrigate with a surface drip as soon as soil moisture decreased to 70% of FC

F1 = Irrigate with surface ooding as soon as soil moisture decreased to 50% of FC

F2 = Irrigate with surface ooding as soon as soil moisture decreased to 60% of FC

F3 = Irrigate with surface ooding as soon as soil moisture decreased to 70% of FC

142 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
ing heavy rainfall. The intermediate daily SWC was estimated using
the following equations:
i The FAO-56 procedure was followed to calculate the daily actual
crop evapotranspiration (Allen et al., 1998).
ETc = Kc ET0 (1)
where Kc is the crop coefcient, and ET0 is the reference evapotran-
spiration.
ii The crop coefcient (Kc ) relationship developed by Gao et al.
(2009) was used to determine the daily crop coefcient through-
out the experimental period.
iii Daily ET0 was calculated according to the Penman-Monteith
equation as recommended by Allen et al. (1998) using the daily
weather data collected at the experimental site.
iv The daily soil water content for each individual treatment was
calculated based on a soil water balance equation used by Hillel
(1998).
S = P + I + U Dw R ETc (2)
where S is the change in soil water storage in the 0100 cm soil
prole, P is precipitation, I is the irrigated water depth, U is the
upward capillary rise from the soil prole below 100 cm, Dw is the
downward drainage beneath the 100 cm soil prole, R is the sur-
face runoff, and ETc is the crop evapotranspiration calculated from
Eq. (1). No runoff (R) occurred, even from the surface ooding treat-
ment, because of the appropriate bund height around the sub-plots;
therefore, the relevant term was neglected when calculating S.
The upward and downward movements of water were estimated
using Darcys law (Gao et al., 2010; Kar et al., 2007).
The rst irrigation event for all treatments was carried out on
the same date when the wheat turned green (Gao et al., 2014;
Zhang et al., 2002), whereas the successive irrigation schedules
were determined on the basis of the soil water content according to
the treatment design explained above. Precise water meters were
connected to each individual subplot. During each irrigation event,
the amount of water for the sprinkler and drip irrigation were set
to 30 mm, whereas 60 mm was used for surface ooding (Li et al.,
2010). The irrigation schedule and total amount of irrigation water
for individual treatments in both cropping seasons are presented
in Table 2.
The soil temperatures in each treatment were measured contin-
uously throughout the experimental period by installing a precise
thermal resistor sensor connected to a data logger. The thermal
resistor sensors were buried in the soil at depths of 0, 5, 10, 15, 20
and 40 cm. The time interval for the data logger (TingmData Center
V6.0, China) to record the temperature was maintained at 30 min.
The data logger battery was replaced every month to prevent any
data from being lost. Before installing the resistor and data logger in
the experimental plots, calibrations were performed for the electric
current, and the temperature was adjusted in a thermostatic water
bath.
To study the morphology of the root system in the soil prole,
each 10 cm of a soil core was collected as a sample from the soil
surface down to a 100 cm depth using a soil auger with an internal
diameter of 6.91 cm and an outer diameter of 7 cm. The samples
were collected from the center of the rows as well as from the
middle section of two adjacent rows of each treatment (Lv et al.,
2010). The core samples were carefully transferred from the auger
to a net mesh bag and kept in the water until all of the soil clods
detached from the roots. The soil was ushed using garden hoses
at low pressure, and all of the clean roots were transferred into a
sieve (0.25 mm2 mesh size) suspended in a trough that was par-
tially lled with water. The live wheat roots (white or pale brown)
were collected from the sieve using tissue forceps and were placed
into transparent plastic cups with water for scanning (Guan et al.,
2015). Root samples were collected twice in 20132014 and 4 times
in 20142015 at different stages of crop growth. To simulate root
water uptake, 15 plant sample measurements were carried out
once a week to determine the plant height, number of tillers per
meter, and leaf area index in each treatment as described by Gao
et al. (2010).
WinRHIZO Reg. 2007d (Regent Instrument Inc.) software was
used to analyze the root morphology, including the root length,
average root diameter, surface area of roots and projected area of
roots based on a scanned image of clean washed roots obtained
using an EPSON PREFECTIONTM V700 Photo Flatbed Scanner-6400
dpi x 9600 dpi (Guan et al., 2015). The root length density (RLD) was
calculated from the obtained root length divided by the sample core
volume (375.013 cm3 ). After the morphological measurement, the
root samples were dried at 70 C for 48 h (Ktterer et al., 1993).
An electronic weighing balance (accuracy of 4 decimal points) was
used to measure the root biomass.
2.5. Estimation of root water uptake and irrigation efciencies
The HYDRUS-1D software package for simulating the one-
dimensional movement of water in variable-saturated media,
developed by Sim unek
et al. (2013), was used as a tool to simu-
late the root water uptake (RWU) and soil water dynamics of the
vadose zone. The effects of the lateral movement of water ow
were ignored. Then, the one dimensional vertical ow in Richards
equation supplemented with the sink term S (z, t) expressed
as the volume of water removed from a unit volume of soil per
unit time due to plant water uptake was used to determine the
RWU without compensation as suggested by Feddes et al. (1978).
Here, the threshold values of h1 , h2 , h3 and h4 for the root water
uptake- water stress response function were taken as 0, 1, 500,
16000 cm, respectively, as suggested by Wesseling et al. (1991).
The initial and boundary conditions are given as follows:
The HYDRUS-1D code was used to simulate RWU from the date
of the rst irrigation event until the time of harvest (i.e., the num-
ber of time-variable boundary records for 81 days in both years).
Three types of soil materials were identied in ve root zones,
each with a depth of 20 cm deepening from the surface through
the 1 m soil prole (presented in Table 1), and were used in the
simulation. The maximum number of iterations and water content
tolerance was retained the same as the default values, i.e., 11 and
0.001, respectively, which was recommended in the software. The
van Genuchten (1980) hydraulic model was preferred with no hys-
teresis, in which Atmospheric BC with Surface Layer was adopted
as an upper boundary condition, and Free Drainage was consid-
ered for the lower boundary condition with the In Water Content
as the initial condition. A maximum ponding depth of 10 cm was
assumed at the surface.
Water productivity (WP) is dened as the ratio of the crop yield
to the water consumed to produce that yield, i.e. evapotranspira-
tion, and irrigation water productivity (IWP) is the increased crop
yield produced per unit amount of irrigation water applied (Ali
et al., 2007; Li et al., 2017) as shown below:
Y
WP = 1000 (3)
ET
Y Yrf
IWP = (4)
I
where, WP or IWP is measured in kg m3 ; ET is evapotranspiration
(mm); and I is the total irrigation amount (m3 /m2 ); Y is the grain
yield (kg/m2 ) in the irrigated plots, and Yrf the average yield (kg/m2 )
of three non-irrigated (rain fed) sample plots in close proximity
to the treatment plots; these sample plots were subjected to the
 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150 143

Table 2
Irrigation scheduling and total irrigation amount for all the treatments in 2013-14 and 2014-15 crop season.

Year Irrigation S1 S2 S3 D1 D2 D3 F1 F2 F3

2015 Schedule 2015/3/18; 3/18; 3/26; 3/18; 3/20; 3/18; 4/23; 3/18; 3/27; 3/18; 3/20; 3/18; 4/17 3/18; 4/16; 3/18;3/23;
4/17; 5/15 4/16; 4/29; 3/24; 4/16; 5/15 4/17; 4/28; 3/27; 4/14; 5/19 4/20; 5/20
5/21 4/28; 5/13; 5/20 4/24; 5/15
5/21
Amount (mm) 90 150 210 90 150 180 120 180 240

2014 Schedule 2014/3/13; 3/13;3/21; 3/13; 3/15; 3/13; 3/25; 3/13; 3/19; 3/13; 3/15; 3/13; 4/7; 3/13;3/28; 3/13;3/17;
3/28; 4/4; 3/28; 4/4; 3/21; 3/28; 4/1; 4/28 4/1; 4/7; 3/20; 3/28; 5/6 4/9; 5/15 4/1; 4/29;
4/28 4/22; 5/4 4/4; 4/22; 4/29; 5/16 4/4; 4/28; 5/15
5/4; 5/16 5/16
Amount (mm) 120 180 240 120 180 210 180 240 300

same environmental conditions and received the same fertilizer

application as in the treatment plots.

2.6. Analysis of simulated output

The simulated output data obtained from the HYDRUS -1D

model were evaluated by calculating the coefcient of determina-
tion (R2 ), root-mean-square error (RMSE), overall index of model
performance (OI), and index of agreement (IA) for the observed and
simulated soil water contents in each treatment (Han et al., 2015;
Legates and McCabe, 1999; Mattar and Alamoud, 2015);
n    2 Fig. 1. Precipitation and atmospheric temperature for crop season 20132014 and
i=1
CO,i C O CS,i C S 2014-15.
2
R = n  2 n  2 (5)
i=1
CO,i C O i=1
CS,i C S
the research site, are presented in Fig. 1 from mid-March to early
 1 n 12 June for both cropping seasons.
RMSE = (CO,i CS,i )2 (6)
n i=1
3.2. Soil water content
n  2
i=1
CO,i CS,i
E =1 n  2 (7) The soil water content (SWC) of all treatments was highest at a
i=1
CO,i C O soil depth of 2050 cm (Fig. 2). In the 70% irrigation treatments

1
 RMSE
OI = 1 +E
2 Cmax Cmin
n  2
CO,i CS,i
IA = 1
i=1
n  2
|CS,i C O | + |CO,i C O |
i=1
where, Cs is the simulated value, Co is the observed value, C S is
the mean of the simulated data, C O is the mean of the observed
data, Cmax is the maximum observed value, Cmin is the minimum
observed value for the soil water content and n is the total number
of observations used in the evaluation. E is the efciency coefcient
and represents an improvement over the coefcient of determina-
tion (R2 ). Analysis of variance (ANOVA) for Two Factorial analysis
in Microsoft Excel was used to determine the statistical signicance
of the treatments at a 5% signicance level ( = 0.05).
3. Results
3.1. Climatic conditions
The total amount of rainfall (Table 7) was less (107.7 mm) in
2014 than in 2015 (137.9 mm), whereas the average atmospheric
temperature was higher in 2014. Thus, the 2014 winter wheat sea-
son was relatively drier than the 2015 season. These factors were
found to play a major role in controlling the irrigation schedule
and total amount of irrigation water applied (Table 2), even for the
same scheduling level and the same irrigation method in the two
corresponding seasons. The precipitation and mean daily atmo-
spheric temperature during the experimental period, recorded at
(IT), the SWC was higher in a deep prole, following the order
(8) F3 > S3 > D3, whereas the SWC in the upper layers was higher in D3,
i.e. D3 > F3 > S3. The SWC below 50 cm in 50% IT was much lower,
i.e., 0.110.16, 0.120.17 and 0.130.18 cm3 /cm3 in SF, SDI and SI,
respectively. The prole water distribution for 60% IT was similar
(9)
for all irrigation methods and was uniformly distributed to a depth
of 80 cm. Below 80 cm, the soil water content was consistently
lower for all irrigation treatments.
3.3. Soil temperature
The irrigation method and scheduling signicantly affected the
surface soil temperature. A maximum surface soil temperature uc-
tuation difference of 4.3 C was found between S1 and S3 when
irrigation scheduling was compared, and the greatest temperature
uctuation that occurred was a difference of 3.7 C between SF and
SDI for the 50% and 60% irrigation methods, respectively (Fig. 3ac).
In the deeper soil prole, the temperature variation was negligible
for both the irrigation method and irrigation schedule (Fig. 3d).
The maximum surface temperature was the highest in S1 (27.3 C)
and the lowest in F3 (24.2 C). The temperature in SF was lower
compared to that of SDI, which agrees with the result presented
by Evett et al. (1995). The maximum and minimum temperature
uctuations in the surface soil were ranked as SDI > SI > SF for the
irrigation method and 50% > 60% > 70% for the irrigation frequency.
The optimum range of surface temperature uctuation in D1 was
6.2 C higher than in S3. In the deeper soil prole (40 cm deep),
the range in temperature variation was less than 1.3 C, with the
highest temperature uctuation in S1 and the lowest in F3.
and Fig. 1
144 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150

Fig. 2. Soil water content (cm3 /cm3 ) in the year 2015.

Fig. 3. Surface soil temperature in (a) sprinkler irrigation (S), (b) surface drip irrigation (D), (c) surface ooding (F) treatments, irrigating at 50% (1), 60% (2) and 70% (3) of FC
and (d) temperature at 40 cm depth for all treatments in 2015 season.
 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150 145

Table 3
Root length density (cm/cm3 ) in 2014 and 2015 measured at different soil depths and at different date (crop stage).

Year Depth Date S1 S2 S3 D1 D2 D3 F1 F2 F3

2015 020 cm 3/18 16.61 0.26e 18.47 0.26bcd 19.16 0.09b 16.80 1.02de 18.69 1.06bc 21.78 0.67a 16.33 2.14e 17.35 0.44cde 21.74 1.53a
4/9 22.10 1.02e 28.55 1.32c 30.94 1.53b 24.58 1.09d 30.66 1.16bc 35.76 0.26a 20.94 2.63e 25.38 0.97d 31.52 0.35b
5/6 28.26 0.73d 32.99 0.94c 40.24 1.33b 28.46 1.01d 34.00 1.37c 44.12 0.92a 28.04 0.18d 29.19 0.52d 40.57 0.06b
6/6 13.92 1.15e 20.71 1.41d 23.97 1.65c 20.11 0.30d 22.07 0.92 cd 29.87 1.24a 13.37 2.30e 20.56 1.64d 27.03 1.66b
2160 cm 3/18 3.19 0.03f 3.56 0.08e 4.09 0.09c 3.30 0.20f 4.47 0.17b 5.11 0.10a 2.63 0.02g 3.87 0.07d 4.64 0.18b
4/9 3.90 0.03f 4.70 0.14d 5.04 0.08c 4.51 0.09e 5.55 0.21b 6.85 0.06a 3.05 0.02g 4.86 0.07 cd 5.65 0.07b
5/6 3.93 0.02g 4.92 0.05e 5.41 0.15c 4.62 0.05f 6.14 0.10b 7.33 0.01a 2.83 0.20h 5.22 0.06d 6.14 0.05b
6/6 3.38 0.03g 4.63 0.05e 5.28 0.06c 3.68 0.13f 5.63 0.08b 6.40 0.06a 1.70 0.24h 4.89 0.03d 5.650.08b
61100 cm 4/9 4.03 0.07b 2.63 0.03d 1.52 0.12e 1.69 0.04e 3.32 0.10c 0.89 0.02f 5.43 0.11a 2.81 0.01d 2.63 0.36d
5/6 9.23 0.26b 7.62 0.13de 4.85 0.10g 7.22 0.03f 8.67 0.23c 3.24 0.06h 11.22 0.18a 7.76 0.40d 7.34 0.35ef
6/6 6.70 0.07b 6.29 0.13c 3.97 0.05f 4.89 0.08e 6.34 0.34c 2.47 0.05g 8.93 0.12a 6.29 0.08c 5.37 0.16d

2014 020 cm 4/30 21.36 0.19d 24.63 0.07c 27.73 0.93b 17.20 0.43e 25.41 1.93c 32.96 0.21a 16.05 0.90e 20.26 0.13d 25.12 0.17c
5/30 10.52 0.37f 15.46 0.09c 16.52 0.49b 12.15 0.92e 16.50 1.08b 22.31 0.73a 7.65 0.25g 14.27 0.34d 16.74 0.20b
2160 cm 4/30 3.71 0.08d 3.89 0.79 cd 3.84 0.09 cd 4.45 0.32bc 5.06 0.54b 8.44 0.08a 2.84 0.10e 3.93 0.02 cd 5.07 0.43b
5/30 3.19 0.04d 3.66 0.42c 3.74 0.17c 3.55 0.23 cd 4.64 0.31b 7.37 0.10a 1.71 0.19e 3.68 0.17c 4.66 0.12b
61100 cm 4/30 4.98 0.59b 2.19 0.11d 0.91 0.18e 1.19 0.12e 3.65 0.03c 0.67 0.34e 7.93 0.55a 2.25 0.03d 2.31 0.43d
5/30 3.62 0.20b 1.81 0.11d 0.75 0.25e 0.81 0.01e 2.67 0.02c 0.51 0.23e 6.31 0.32a 1.83 0.05d 1.69 0.04d
Letters indicate statistical signicance at = 0.05 level within the same row with a, b, c and so on showing the statistical difference from the highest to the lowest and the
values are the means of three replication.

Table 4
Root morphology at owering stage in 2014 and 2015.

Year Parameters S1 S2 S3 D1 D2 D3 F1 F2 F3
2015 Total Projected area (m2 /m3 ) 12.8 0.26f 15.6 0.36d 17.3 0.53c 12.8 0.11f 16.6 0.40c 21.0 0.05a 12.9 0.68f 13.9 0.26e 18.3 0.63b
Total Surface area (m2 /m3 ) 40.1 1.02g 49.0 1.14e 54.5 1.40c 40.1 0.48g 52.3 0.86d 65.9 0.18a 40.5 1.01g 43.8 0.40f 57.5 0.95b
Average Diameter (mm) 0.24 0.004c 0.28 0.01ab 0.28 0.01ab 0.29 0.01ab 0.26 0.01bc 0.29 0.01b 0.30 0.003a 0.27 0.23b 0.27 0.01b

2014 Total Projected area (m2 /m3 ) 10.8 0.66d 11.6 0.91d 12.2 1.33 cd 8.08 1.11e 15.2 0.59b 17.0 1.0a 7.6 0.37e 11.3 0.35d 13.1 0.45c
Total Surface area (m2 /m3 ) 34.0 0.93g 36.6 1.16b 38.4 1.46g 25.4 1.19ef 47.9 0.87e 53.4 1.07f 23.9 0.41c 35.5 0.65a 41.1 0.60d
Average Diameter (mm) 0.23 0.01ab 0.21 0.03b 0.21 0.01b 0.23 0.01b 0.26 0.01a 0.23 0.00b 0.218 0.01b 0.23 0.01b 0.23 0.01ab
Letters indicate statistical signicance at = 0.05 level within the same row with a, b, c and so on showing the statistical difference from the highest to the lowest and the
values are the means of three replication.

Table 5
Dry root biomass (g/m2 ) at owering stage in 2014 and 2015 at different soil prole.

Year Depth (cm) S1 S2 S3 D1 D2 D3 F1 F2 F3

2015 020 298.4 0.75g 371.3 0.85d 387.4 1.27c 362.7 1.01f 372.5 1.30d 491.7 0.93a 252.3 0.24h 364.8 0.45e 465.3 0.17b
2060 22.61 0.02h 23.76 0.08f 29.07 0.17d 23.39 0.07g 29.89 0.15c 49.60 0.02a 22.05 0.25i 28.59 0.16e 43.41 0.09b
60100 65.81 0.48b 43.65 0.28e 40.11 0.11f 40.61 0.07f 49.97 0.15c 32.83 0.09g 154.10 1.3a 46.48 0.65d 43.65 0.85e

2014 020 234.9 0.63e 235.4 0.47e 304.4 1.57b 174.3 0.71g 281.2 2.1c 315.7 0.88a 154.8 0.75h 226.1 0.40f 263.7 0.26d
2060 17.69 0.07g 18.58 0.98f 18.4 0.11fg 25.47 0.39d 28.26 0.49c 32.57 0.13a 11.03 0.18h 21.29 0.06e 31.01 0.50b
60100 17.80 0.63b 7.36 0.09f 5.20 0.40g 5.42 0.16g 15.70 0.08c 3.22 0.36h 39.64 0.77a 8.36 0.04e 13.25 0.33d
Letters indicate statistical signicance at = 0.05 level within the same row with a, b, c and so on showing the statistical difference from the highest to the lowest and the
values are the means of three replication.

Table 6
Statistical parameters computed to evaluate accuracy in simulation for 2014 and 2015.

Year Statistical Parameters S1 S2 S3 D1 D2 D3 F1 F2 F3

2015 Coefcient of determination (R2 ) 0.99 0.75 0.65 0.94 0.77 0.64 0.83 0.76 0.64
Root mean square error (RMSE) 0.01 0.04 0.05 0.01 0.04 0.06 0.04 0.05 0.07
Overall index of model performance (OI) 0.98 0.77 0.63 0.91 0.81 0.7 0.79 0.79 0.71
Index of agreement (IA) 0.99 0.92 0.89 0.98 0.93 0.88 0.92 0.93 0.88

2014 Coefcient of determination (R2 ) 0.77 0.98 0.89 0.92 0.8 0.77 0.64 0.81 0.51
Root mean square error (RMSE) 0.02 0.04 0.02 0.03 0.04 0.02 0.03 0.04 0.06
Overall index of model performance (OI) 0.71 0.96 0.75 0.66 0.81 0.82 0.72 0.79 0.60
Index of agreement (IA) 0.80 0.99 0.85 0.69 0.92 0.93 0.85 0.91 0.82

Table 7
Soil water balance in the root zone for crop season 2014 and 2015.

Year Parameters S1 S2 S3 D1 D2 D3 F1 F2 F3

2015 Total Rainfall (R), mm 138 138 138 138 138 138 138 138 138
Total Irrigation (I), mm 90 150 210 90 150 180 120 180 240
Change in Soil Water Storage (S), mm 30 11 16 38 22 11 42 17 9
Drainage (D), mm 5 8 25 9 10 19 14 37 49
Evapotranspiration (ETc ), mm 263 291 307 275 300 310 286 298 319
Actual Root Water Uptake (TA ), mm 190 226 256 210 248 265 238 256 288

2014 Total Rainfall (R), mm 108 108 108 108 108 108 108 108 108
Total Irrigation (I), mm 120 180 240 120 180 210 180 240 300
Change in Soil Water Storage (S), mm 41 21 10 30 22 10 16 1 5
Drainage (D), mm 9 28 36 4 19 29 30 46 62
Evapotranspiration (ETc ), mm 259 281 302 254 291 299 273 303 341
Actual Root Water Uptake (TA ), mm 166 197 227 164 209 225 190 228 282
146 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
3.4. Root morphology
The root morphology parameters (length, projected area, sur-
face area, and diameter) determined using the WinRHIZO software
and dry root biomass is presented in the following sections. The
detailed root length density (RLD) data for the 020, 2060, and
60100 cm soil depth, measured on different dates during both
cropping seasons (Table 3), revealed that maximum root growth
occurred during the owering stage. In 2015, the D3 treatment
resulted in denser roots (35.76 cm/cm3 ) in the early stage, followed
by F3 (31.52 cm/cm3 ) and S3 (30.94 cm/cm3 ) and reached a maxi-
mum RLD of 44.12 cm/cm3 at owering, which was followed by the
same treatments, i.e., F3 (40.57 cm/cm3 ) and S3 (40.24 cm/cm3 ).
The root length density (RLD) distributions at owering are shown
in Table 3 (April 30th and May 6th in 2014 and 2015 respectively).
The SDI also enhanced the growth rate in the upper soil prole com-
pared with SF and SI, but lower RLD was measured below 60 cm.
After the booting stage, the highest root growth rate was observed
in the SI and SF treatments, but did not compete with the total RLD
produced by SDI for different schedules in top soil. These results
are similar to those presented by Yao (2005) and Yang et al. (2006).
The RLD below 60 cm was the highest in F1 (11.22 cm/cm3 ) and fol-
lowed the order SF > SI > SDI and 50% > 60% > 70% for the irrigation
schedules except for SDI at 60% (D2, 8.67 cm/cm3 ), which resulted
in greater RLD than F2 (7.76 cm/cm3 ) and S2 (7.62 cm/cm3 ). The
prole RLD pattern at the owering stage in 2014 was quite similar
to that in 2015 (Table 3) for a particular soil prole; however the
reverse order was observed when comparing RLD at 020 cm and
60100 cm for both seasons.
The average diameter of the roots sampled during owering
at a 1 m depth (Table 4) was the highest in D2 (0.255 mm) fol-
lowed by S1 (0.232 mm) in 2014 versus F1 (0.303 mm) followed
by D1 (0.287 mm) in 2015. The lowest diameter was found in S2
(0.210 mm) and S1 (0.241 mm) for the corresponding years. The
results were not similar between the two years and did not follow
any xed order for either season. However, a thicker diameter was
associated with the low frequency irrigation treatment, and a thin-
ner diameter was observed in the treatments with high amounts
of applied water. The highest values for the total root surface area
(SA) were measured in the D3 treatment, i.e., 53.4 and 65.9 m2 /m3
the 2014 and 2015 cropping seasons, respectively, and the lowest
values were 23.9 and 40.1 m2 /m3 , measured in F1 and D1, respec-
tively (Table 4). A similar pattern was observed for the root length
density. The same order was observed for the projected area (PA)
and the root surface area.
The dry root biomass in the soil prole was measured in both
seasons at 10 cm intervals up to 100 cm. The biomass was cate-
gorized into three zones, i.e., depths of 020 cm, 2060 cm and
60100 cm, as shown in Table 5. The highest root biomasses,
491.7 g/m2 and 315.7 g/m2 , were measured during the owering
stage in D3 (020 cm) in 2015 and 2014, respectively, followed by
F3 (465.3 g/m2 ) in 2015 and S3 (304.4 g/m2 ) in 2014. The lowest
root biomasses observed were 252.3 g/m2 and 154.8 g/m2 in the F1
treatment in 2015 and 2014, respectively. Signicant differences in
the dry root biomass were observed between the treatments under
different irrigation schedules ( = 0.01) and irrigation methods
( = 0.05) when evaluating the three rooting zones, i.e., 020 cm,
2060 cm and 60100 cm, separately. In 2015, the dry mass pattern
in the top 60 cm decreased with decreasing irrigation frequency for
all irrigation methods, lead by SDI followed by SI, except at 70%
where SF exceeded SI. Compared with the top 60 cm, the reverse
order was observed for the root biomass pattern at the 60100 cm
soil depth, which was very similar to the pattern observed for RLD
(Table 3) at respective depths during the owering stage. The pat-
tern observed for the different irrigation schedule agreed with the
results presented by Ktterer et al. (1993), and the pattern for the
different irrigation methods was similar to the results reported by
Bai and Li (2003). The pattern for biomass in 2014 differed slightly
from that observed in 2015 for the corresponding depths because
of different weather conditions (especially rainfall) and soil tem-
perature variations during the different growing seasons.
3.5. Data verication
The simulated soil water content obtained from the HYDRUS
model run was evaluated against the actual measured soil water
content obtained in the eld by computing R2 , RMSE, OI and IA for
each treatment; the data are presented in Table 6 for both cropping
seasons. Evaluating the simulated SWC on the basis of the index
of agreement (IA) conrmed that the overall performance of the
simulated results was accurate. The IA value ranged from 0.88 to
0.99 in 2015 and from 0.8 to 0.99 in 2014, showing that the mea-
sured and simulated SWC were matching and hence the simulated
output from HYDRUS was sufciently accurate. Similarly, a lower
RMSE (0.01 0.04) and values closer to 1 for R2 and OI in treatments
that were irrigated at 50% and 60% of FC showed good agreement
between the simulated and observed SWC. However, the values of
these parameters (RMSE, R2 and OI) for the treatments that were
irrigated to 70% of FC showed a relatively lower agreement, which
occurred because of excessive SWC. Excessive moisture on the sur-
face in the 70% treatments resulted in higher ET and drainage, which
caused rapid depletion of soil water and hence reduced the accuracy
of the measurements.
3.6. Soil water balance
3.6.1. Root water uptake (RWU)
The daily RWU simulated by HYDRUS-1D (version 4.17) is
shown in Fig. 4 (a, b, and c) for 2015. The daily simulation graph
is not shown for 2014, but the total RWU from re-greening to har-
vest is presented in Table 7 for both years. The simulation results
showed that the daily RWU reached approximately 69 mm/d at
the end of March to mid-April and during early to mid-May (Fig. 4).
Synchronizing the daily uptake with the rainfall pattern (Fig. 1)
and irrigation scheduling (Table 2) showed that the RWU cor-
respondingly increased with increased soil water availability at
grain lling. Similar results were found by Li et al. (2014) and Xue
et al. (2003), who reported higher RWU after rainfall or irrigation
and at the grain-lling stage of maize. Statistical analysis showed
that the method used to apply the irrigation water signicantly
(=0.05) affected the total RWU from re-greening to harvest; the
differences were highly signicant (=0.01) for different irrigation
schedules. The total uptake in both years was higher for the 70%
treatment schedule and decreased with less frequent amounts of
irrigation (Table 7). The simulation showed that F3 used 60 mm and
90 mm more irrigation water to increase the RWU by 22.6 mm and
57.1 mm, respectively, compared with D3, and F3 needed 30 mm
and 60 mm more irrigation to increase the RWU by 31.5 mm and
54.8 mm compared to S3 in 2015 and 2014, respectively. This result
shows that when winter wheat was irrigated at a high frequency,
more efcient utilization of the applied irrigation water occurred
in the SDI system than in SI and SF because SDI had a higher RLD in
the top soil. Since the SDI system reduced the wetting of the sur-
face soil, considerable amount of irrigation water could be saved
from surface soil evaporation, which could signicantly reduce
non-benecial water consumption. A total of 30 mm and 60 mm
more irrigation water was used in the F1 treatment in 2015 and
2014, respectively, to increase RWU by 28.04 mm and 26.7 mm,
compared to D1, and to increase the RWU by 47.4 mm and 24.8 mm
in 2015 and 2014, respectively, compared to S1. These results indi-
cate that the RWU was enhanced by RLD under surface ooding.
Additionally, the results suggest that the irrigation schedule and
 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
Fig. 4. Daily root water uptake (RWU) from (a) sprinkler irrigation (S), (b) surface
drip irrigation (D), and (c) surface ooding (F) treatments, irrigated at 50% (1), 60%
(2) and 70% (3) of FC in 2015.
water application method play more signicant roles, generating
hydraulic head in the root zone depending on the weather and soil
conditions.
Analysis of the simulated prole for the RWU distribution
obtained for each soil depth within the 1 m root zone revealed that
maximum water uptake occurred in the top 20 cm soil layer in F3
(10.84 cm in 2015 and 10.87 cm in 2014), followed by D3 (10.33 cm)
in 2015 and F2 (8.91 cm) in 2014 (Fig. 5). The result also demon-
strated that the top 20 cm soil layer was the main uptake region,
which contributed approximately 38%40% of the total RWU. The
uptake from the sub regions from 20 cm to 60 cm and 60 cm to
100 cm ranged from 49% to 51% and 10% to 12% of the total RWU,
respectively. The prole distribution of the RWU shows that uptake
from 60 to 100 cm was much lower compared to that from 20 to
60 cm. On the other hand, the high frequency irrigation treatments
consistently reduced the uptake rate in the deeper zone, whereas
the low frequency treatment increased the rate of uptake in the
deeper prole. Therefore, the RWU shifted downward as the soil
water content decreased in the top prole.
3.6.2. Evapotranspiration (ET)
In both cropping seasons, a signicant ( = 0 0.05) difference
in evapotranspiration (ET) was found for the different irrigation
methods and schedules. The total ET for individual treatments did
not vary signicantly between seasons when accounting for the
147
Fig. 5. Prole root water uptake (RWU) distribution in (a) 2014 and (b) 2015 for all
irrigation treatments.
seasonal irrigation amount. The maximum ET of 341 mm in 2014
and 319 mm in 2015 was measured in the F3 treatment (Table 7).
Soil evaporation (E) was higher in 2014 because the number of
irrigation events per treatment increased with less precipitation
(Table 2, Fig. 1). The seasonal ET increased with increasing irrigation
amount, as shown by the 70% IT, and found to be more in surface
ooding compared to other irrigation methods. The evaporation
(Es) to evapotranspiration (ET) ratio decreased with increasing irri-
gation amounts, as described by Yu et al. (2009). The grain yield
(Y) increased linearly (Y = 0.018 ET + 3.404) with increasing ET in
2015. This outcome showed that the slope of the linear line for the
relationship between yield and ET was estimated to be 1.8 kg/m3 ,
which was approximately the same value (1.73 kg/m3 ) presented
by Zhang et al. (1999) for winter wheat in the piedmont region of
the NCP.
3.7. Grain yield and irrigation efciency
The grain yield (GY), water productivity (WP) and irrigation
water productivity (IWP) estimated from a harvested 1 m2 sample
per treatment are presented in Fig. 6. The analysis showed that irri-
gating winter wheat with different irrigation methods signicantly
( = 0.05) inuenced the grain yield, and the differences were found
to be highly signicant ( = 0.01) for different irrigation schedules.
The yield data for this study varied from 9.53 t/ha in D2 to 8.26 t/ha
in F1, which supported the simulated potential yield presented by
Lu and Fan (2013), in which a potential yield of 9.1 t/ha was found
for winter wheat in the NCP. These authors further illustrated that
yield gap of 2.7 t/ha to the potential yield can be recovered through
proper irrigation management. This could easily be obtained for
irrigation to 60% of FC, and a maximum yield for SDI followed by
SI occurred at all irrigation levels, except at 50%. Irrigating at a low
SWC (50% of FC) drastically reduced the yield. Surface drip irriga-
tion not only increased the yield but also increased the WP. The
148 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
Fig. 6. Grain yield (GY), water productivity (WP) and irrigation water productivity
(IWP) in 20142015 crop seasons.
highest WP value of 2.08 kg/m3 was measured in the D2 treatment,
followed by S2 (2.05 kg/m3 ), and was the lowest in F1 (1.86 kg/m3 ).
The WP value found in this study was similar to values measured
by Lv et al. (2013b) and Wang et al. (2015). They found that WUE
ranged from 1.7 to 4.2 kg/m3 and 1.9 to 2.06 kg/m3 . The most irri-
gated treatments, i.e., F3, S3, F2 and D3, lowered the IWP, with the
lowest value for F3 (2.81 kg/m3 ), whereas the low irrigation treat-
ments, S1, D1, F1, D2 and S2, exhibited higher IWP values, with the
highest value in S1 (6.89 kg/m3 ), followed by D1 (6.71 kg/m3 ).
4. Discussion
4.1. Prole soil water content
The soil water content in a prole depends on the water bal-
ance between the inow and outow of the prole. Because the
climatic conditions and soil characteristics were the same for all
treatments, the prole water distribution differed because of the
different irrigation water application methods and irrigation sched-
ules. The soil permeability of the central root zone (2060 cm) was
found to be comparatively low (Table 1), which may be one of the
reasons that the root zone maintained a higher SWC in most of
the treatments. Similarly, soil permeability below 80 cm was very
high, which allowed for free drainage and kept the SWC at a low
level. However, the SWC in the F3, S3 and D3 treatments was higher
because of frequent irrigation and increases in the hydraulic head.
The results show that the total irrigation water required for supe-
rior growth and development of winter wheat can be managed in
the top 60 cm of the soil prole, and wetting of the lower soil pro-
le below 60 cm could be restricted by applying a lower amount of
water per irrigation event under proper irrigation scheduling with
suitable irrigation methods. Similar results were found by Rawlins
and Raats (1975).
4.2. Soil temperature uctuation
The temperature of the surface soil uctuated signicantly
because of the very low SWC, which allowed the temperature to
increase rapidly with less solar radiation compared to the temper-
ature uctuation in the deeper soil prole. The soil temperature
was higher in SDI and lower in SF, which was previously explained
by Evett et al. (1995), who predicted warmer soils under SDI com-
pared to surface irrigation. From these results, we concluded that
the irrigation schedule greatly affects the top soil temperature, and
this effect may occur due to different vegetative growth as well as
different wetting patterns of the surface soil under different irriga-
tion schedules. Deeper in the soil prole, the temperature variation
was minimal under different irrigation methods and schedules.
The higher frequency of the irrigated treatments D3, F3 and S3
maintained the top soil at a lower temperature than the low-
frequency irrigation treatments D1, F1 and S1 because of the longer
duration of wet soil, which would conduct solar heat, as well as
higher canopy coverage. The temperature variation in the soil pro-
le was similar to the range reported by Lv et al. (2013a).
4.3. Root development
4.3.1. Root length density distribution
In SDI, a low amount (30 mm) of water was applied during
each irrigation event with relatively long application times. Con-
sequently, the upper 60 cm of the soil prole held almost all of the
irrigation water, causing the RLD to shift upward compared to the
other irrigation methods. Under surface ooding, large amounts
of water were applied at once, and an increase in piezometric head
forced the moisture to move downward, which encouraged the root
system moving downward (Li et al., 2010). In contrast, when irrigat-
ing winter wheat at 70% of FC, the SWC in the top 60 cm consistently
remained at a higher level and may have been associated with less
aeration, which caused the root systems to remain in the upper
20 cm (Zhang et al., 2009). The irrigation interval in the 50% irri-
gation treatment was longer, and therefore, the root system failed
to fulll the water requirements of the crop in the upper prole
because the soil water depleted very rapidly in the top soil com-
pared to the lower soil prole, forcing the roots to grow downwards.
Thus, we concluded from the results that dense root systems were
present in the upper layer in the 70% irrigation treatment and in the
lower prole in the 50% irrigation treatment. Similar results were
found by Li et al. (2010).
4.3.2. Root biomass variation
The result showed that the roots rapidly colonized the top pro-
le (020 cm) in the frequently irrigated treatments (irrigated to
70% of FC) and produced signicantly more biomass than the less
frequently irrigated treatments (50% of FC). This outcome may be
due to the soil surface temperature variation (Fig. 3) under different
irrigation schedules, which was described by Vincent and Gregory
(1989). The biomass measured in F1 was the lowest at a depth of
060 cm, which also provided evidence that a decit in soil water
under a relatively higher temperature resulted in a signicantly
reduction in dry root biomass (Torreano and Morris, 1998).
4.4. Root water uptake pattern
Maximum root water absorption occurred after full vegetative
growth to the booting stage, from the end of March to mid-April,
and during heading to anthesis, from early to mid-May (Liu et al.,
2002), as shown in Fig. 4. A comparison of the ratio of actual root
water uptake to total water inow (R + I) in the root zone (Table 7)
showed that in 2015, this ratio was the highest in F1, followed by
D1, and the lowest value was measured in S3. In 2014, the high-
est ratio was measured in D2 followed by S1, and the lowest value
was measured in the S2 treatment. In both years, the water uptake
rate was found to be comparatively high in SDI, which had a higher
RLD in the 020 cm soil prole, as described by Zhang et al. (2004).
On the other hand, the surface temperature in SDI and the low fre-
quency irrigation treatments remained higher (Fig. 3), which likely
occurred due to prolonged surface evaporation that reduced the
SWC in the surface soil (Fig. 2). Under this condition, if the lower soil
proles contain sufcient water, the soil moisture gradient estab-
lished between the layers promotes a capillary rise at night when
the surface temperature is low (Li et al., 2010; Lv et al., 2013a). This
phenomenon contributed to maximize the root water uptake and
overall transpiration for the treatments with shallow roots, such
as SDI. By contrast, deep roots develop under ood irrigation or
in low frequency irrigation treatments, and water uptake directly
 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
occurred in the deep soil prole where the RLD distribution plays
an important role in contributing to RWU (Lv et al., 2010; Zhang
et al., 2009).
The main uptake zone established in the top 20 cm of the soil
prole was because of higher RLD in this region. The higher water
uptake from 20 to 60 cm soil prole compared with 60100 cm
resulted from the high available soil water content. This result
showed that the availability of water for crop water use was the
main factor responsible for optimum RWU. The prole RWU in
the irrigation treatments with different water application methods
converged as the depth increased (Fig. 5). This outcome indicated
that the prole water uptake was reduced even in treatments that
were frequently irrigated and that received more irrigation water.
However, the uptake was higher if there were dense roots in the
deep soil prole, which were found in treatments with less frequent
irrigation.
4.5. Evaluation of irrigation water productivity
The irrigation water productivity decreased with increasing irri-
gation amount because the grain yield did not increase linearly
(Li et al., 1995), whereas ET was found to increase linearly with
increasing amounts of irrigation. Irrigating winter wheat to 50%,
60% and 70% of FC using SDI increased the IWP by 33.9%, 28.1%
and 40.8% compared to SF, whereas SI increased the IWP by 37.4%,
22.4% and 16.6%, respectively, compared to SF under the same irri-
gation schedule. A quadratic relationship (Y = 0.00011 I2 + 0.04
 (Y) and total irri-
I + 5.564) was established by plotting grain yield
gation amount (I). Maximizing the function Y = 0 obtained for
I
grain yield indicated that 180 mm of irrigation water (approx-
imately 314 mm, including rainfall) is the optimal requirement
to produce the highest yield. Liu et al. (2011) also calculated
185 mm and 186 mm of irrigation water in the rst and second
seasons, respectively of their study to obtain maximum grain yield
under a sprinkler irrigation system. Similarly, a quadratic function
(WP = 0.000018 I2 + 0.5563 I + 0.348) was found by plotting WP
(equation 3) versus total irrigation amount (I). The maximum WP
was estimated based on an irrigation amount of 154 mm  (approxi-
(WP)
mately 292 mm, including rainfall) when maximizing =0
I usage in northern China: dependence on palaeowaters and effects on water
the relationship obtained for WP. The irrigation amounts for the
D2 and S2 treatments were measured to be 150 mm (Table 7).
This suggests that the irrigation amounts applied in these treat-
ments were sufcient to achieve maximum WP; in fact, the WP in
D2 (2.08 kg/m3 ) was slightly higher than that in S2 (2.05 kg/m3 )
because of the higher grain yield in D2 (Fig. 6). This result shows
that using SDI as soon as the soil water content reaches 60% of the
eld capacity produced the maximum grain yield and highest WP
and also improved the IWP.
5. Conclusion
We concluded that the soil water content, root growth and
soil temperature, which varied according to the water application
methods and irrigation scheduling, play an important role in root
water uptake (RWU) and can affect the overall soil water dynam-
ics of winter wheat cultivation. It is very important to understand
how to irrigate, how much to irrigate and when to irrigate winter
wheat depending on the weather conditions to optimize the RWU.
Approximately 90% of total water uptake occurs in the 060 cm soil
prole, where most of the roots exist. The wide differences in water
uptake rates in the top soil between the low and high irrigation
treatments became narrower in the deeper soil prole. Thus, we
concluded that applying more irrigation water than can be stored
in the top 60 cm of the soil prole will result in inefcient utilization
of this water by winter wheat. Irrespective of the weather pattern,
149
if the soil conditions are similar to those observed in this study, a
surface drip irrigation system scheduled to irrigate at 60% of eld
capacity (with 30 mm of water per irrigation event) will not only
maintain the highest WP with a high RWU but will also improve
the irrigation water productivity and produce higher grain yield.
Acknowledgements
We are grateful for the nancial support from the China Agri-
cultural Research System (CARS-3-1-30), the Special Fund for Agro-
Scientic Research in the Public Interest (201203077, 201501017),
the Chinese National Natural Science Fund (51309227), and the
Agricultural Science and Technology Innovation Program (ASTIP),
Chinese Academy of Agricultural Sciences.
References
unek,
Sim J., Hopmans, J.W., 2009. Modeling compensated root water and nutrient
uptake. Ecol. Modell. 220, 505521.
Simunek,
J., Sejna, M., Saito, H., Sakai, M., van Genuchten, M.T., 2013. The
HYDRUS-1D Software Package for Simulating the Movement of Water, Heat,
and Multiple Solutes in Variably Saturated Media, Version 4.17, HYDRUS
Software Series 3. Department of Environmental Sciences, University of
California Riverside, Riverside, California, USA, pp. 342.
Albasha, R., Mailhol, J.-C., Cheviron, B., 2015. Compensatory uptake functions in
empirical macroscopic root water uptake models experimental and
numerical analysis. Agric. Water Manage. 155, 2239.
Ali, M.H., Hoque, M.R., Hassan, A.A., Khair, A., 2007. Effects of decit irrigation on
yield, water productivity, and economic returns of wheat. Agric. Water
Manage. 92, 151161.
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration:
Guidelines for Computing Crop Water Requirements. United Nations FAO
Irrigation and Drainage Paper No. 56.
Bai, W.-M., Li, L.-H., 2003. Effect of irrigation methods and quota on root water
uptake and biomass of alfalfa in the Wulanbuhe sandy region of China. Agric.
Water Manage. 62, 139148.
Camposeo, S., Rubino, P., 2003. Effect of irrigation frequency on root water uptake
in sugar beet. Plant Soil 253, 301309.
Carvalho, P., Foulkes, M.J., 2013. Roots root and Uptake of Water and Nutrients
roots uptake of water and nutrients. In: Christou, P., Savin, R., Costa-Pierce, B.,
Misztal, I., Whitelaw, C.B. (Eds.), Sustainable Food Production. Springer, New
York, pp. 13901404.
Coelho, E., Or, D., 1999. Root distribution and water uptake patterns of corn under
surface and subsurface drip irrigation. Plant Soil 206, 123136.
Currell, M.J., Han, D., Chen, Z., Cartwright, I., 2012. Sustainability of groundwater
quality, quantity and ecosystem health. Hydrol. Processes 26, 40504066.
Du, T., Kang, S., Zhang, J., Davies, W.J., 2015. Decit irrigation and sustainable
water-resource strategies in agriculture for Chinas food security. J. Exp. Bot.
66, 22532269.
Eissenstat, D.M., 1992. Costs and benets of constructing roots of small diameter. J.
Plant Nutr. 15, 763782.
Ephrath, J.E., Silberbush, M., Berliner, P.R., 1999. Calibration of minirhizotron
readings against root length density data obtained from soil cores. Plant Soil
209, 201208.
Evett, S.R., Howell, T.A., Schneider, A.D., 1995. Energy and water balances for
surface and subsurface drip irrigated corn. Fifth International Microirrigation
Congress.
Fabio, A., Madeira, M., Steen, E., Ktterer, T., Ribeiro, C., Arajo, C., 1995.
Development of root biomass in an Eucalyptus globulus plantation under
different water and nutrient regimes. Plant Soil 168169, 215223.
Feddes, R.A., Kowalik, P.J., Zaradny, H., 1978. Simulation of Field Water Use and
Crop Yield. Pudoc, Wageningen.
Gao, Y., Duan, A., Sun, J., Li, F., Liu, Z., Liu, H., Liu, Z., 2009. Crop coefcient and
water-use efciency of winter wheat/spring maize strip intercropping. Field
Crop. Res. 111, 6573.
Gao, Y., Duan, A., Qiu, X., Liu, Z., Sun, J., Zhang, J., Wang, H., 2010. Distribution of
roots and root length density in a maize/soybean strip intercropping system.
Agric. Water Manage. 98, 199212.
Gao, Y., Yang, L., Shen, X., Li, X., Sun, J., Duan, A., Wu, L., 2014. Winter wheat with
subsurface drip irrigation (SDI) Crop coefcients, water-use estimates, and
effects of SDI on grain yield and water use efciency. Agric. Water Manage.
146, 110.
Guan, D., Zhang, Y., Al-Kaisi, M.M., Wang, Q., Zhang, M., Li, Z., 2015. Tillage
practices effect on root distribution and water use efciency of winter wheat
under rain-fed condition in the North China Plain. Soil Tillage Res. 146 (Part B),
286295.
Han, M., Zhao, C., Feng, G., Yan, Y., Sheng, Y., 2015. Evaluating the effects of mulch
and irrigation amount on soil water distribution and root zone water balance
using HYDRUS-2D. Water 7, 2622.
150 S.K. Jha et al. / Agricultural Water Management 182 (2017) 139150
Hillel, D., 1998. Environmental Soil Physics: Fundamentals, Applications, and
Environmental Considerations. Academic press.
Jin, H., Qingjie, W., Hongwen, L., Lijin, L., Huanwen, G., 2009. Effect of alternative
tillage and residue cover on yield and water use efciency in annual double
cropping system in North China Plain. Soil Tillage Res. 104, 198205.
Ktterer, T., Hansson, A.-C., Andrn, O., 1993. Wheat root biomass and nitrogen
dynamicseffects of daily irrigation and fertilization. Plant Soil 151, 2130.
Kar, G., Kumar, A., Martha, M., 2007. Water use efciency and crop coefcients of
dry season oilseed crops. Agric. Water Manage. 87, 7382.
Kaspar, T.C., Bland, W.L., 1992. Soil temperature and root growth. Soil Sci. 154,
290299.
Kmoch, H.G., Ramig, R.E., Fox, R.L., Koehler, F.E., 1957. Root development of winter
wheat as inuenced by soil moisture and nitrogen fertilization. Agron. J. 49,
2026.
Kramer, P.J., Boyer, J.S., 1995. Water Relations of Plants and Soils. Academic press.
Legates, D.R., McCabe, G.J., 1999. Evaluating the use of goodness-of-t Measures in
hydrologic and hydroclimatic model validation. Water Resour. Res. 35,
233241.
Li, F., Zhao, S., Duan, S., 1995. The strategy for limited irrigation of spring wheat in
semiarid Loess Plateau, China. Chin. J. Appl. Ecol 6, 259264.
Li, Q., Dong, B., Qiao, Y., Liu, M., Zhang, J., 2010. Root growth, available soil water,
and water-use efciency of winter wheat under different irrigation regimes
applied at different growth stages in North China. Agric. Water Manage. 97,
16761682.
Li, C.X., Zhou, X.G., Sun, J.S., Wang, H.Z., Gao, Y., 2014. Dynamics of root water
uptake and water use efciency under alternate partial root-zone irrigation.
Desalin. Water Treat. 52, 28052810.
Li, X., Tong, L., Niu, J., Kang, S., Du, T., Li, S., Ding, R., 2017. Spatio-temporal
distribution of irrigation water productivity and its driving factors for cereal
crops in Hexi Corridor, Northwest China. Agric. Water Manage. 179, 5563.
Liu, C., Zhang, X., Zhang, Y., 2002. Determination of daily evaporation and
evapotranspiration of winter wheat and maize by large-scale weighing
lysimeter and micro-lysimeter. Agric. For. Meteorol. 111, 109120.
Liu, H., Yu, L., Luo, Y., Wang, X., Huang, G., 2011. Responses of winter wheat
(Triticum aestivum L.) evapotranspiration and yield to sprinkler irrigation
regimes. Agric. Water Manage. 98, 483492.
Lu, C., Fan, L., 2013. Winter wheat yield potentials and yield gaps in the North
China Plain. Field Crop. Res. 143, 98105.
Lv, G., Kang, Y., Li, L., Wan, S., 2010. Effect of irrigation methods on root
development and prole soil water uptake in winter wheat. Irrig. Sci. 28,
387398.
Lv, G., Hu, W., Kang, Y., Liu, B., Li, L., Song, J., 2013a. Root water uptake model
considering soil temperature. J. Hydrol. Eng. 18, 394400.
Lv, L., Yao, Y., Zhang, L., Dong, Z., Jia, X., Liang, S., Ji, J., 2013b. Winter wheat grain
yield and its components in the North China Plain: irrigation management
cultivation, and climate. Chil. J. Agric. Res. 73, 233242.
Mattar, M.A., Alamoud, A.I., 2015. Articial neural networks for estimating the
hydraulic performance of labyrinth-channel emitters. Comput. Electron. Agric.
114, 189201.
McMichael, B., Taylor, H., 1987. Applications and limitations of rhizotrons and
minirhizotrons. Minirhizotron Observation Tubes: Methods and Applications
for Measuring Rhizosphere Dynamics. In. HM Taylor, H.M. (Ed.), 113.
Rawlins, S.L., Raats, P.A.C., 1975. Prospects for high-frequency irrigation. Science
188, 604610.
Ritchie, J.T., 1981. Water dynamics in the soil-plant-atmosphere system. Plant Soil
58, 8196.
Roose, T., Fowler, A.C., 2004. A model for water uptake by plant roots. J. Theor. Biol.
228, 155171.
Rowse, H.R., 1974. The effect of irrigation on the length, weight, and diameter of
lettuce roots. Plant Soil 40, 381391.
Samson, B., Sinclair, T., 1994. Soil core and minirhizotron comparison for the
determination of root length density. Plant Soil 161, 225232.
Shao, L., Zhang, X., Chen, S., Sun, H., Wang, Z., 2009. Effects of irrigation frequency
under limited irrigation on root water uptake, yield and water use efciency of
winter wheat. Irrig. Drain. 58, 393405.
Shao, L.W., Zhang, X.Y., Sun, H.Y., Chen, S.Y., Wang, Y.M., 2011. Yield and water use
response of winter wheat to winter irrigation in the North China Plain. J. Soil
Water Conserv. 66, 104113.
Sun, H.-Y., Liu, C.-M., Zhang, X.-Y., Shen, Y.-J., Zhang, Y.-Q., 2006. Effects of
irrigation on water balance, yield and WUE of winter wheat in the North China
Plain. Agric. Water Manage. 85, 211218.
Torreano, S.J., Morris, L.A., 1998. Loblolly pine root growth and distribution under
water stress. Soil Sci. Soc. Am. J. 62, 818827.
Vincent, C.D., Gregory, P.J., 1989. Effects of temperature on the development and
growth of winter wheat roots. Plant Soil 119, 99110.
Wang, X., Huang, G., Yang, J., Huang, Q., Liu, H., Yu, L., 2015. An assessment of
irrigation practices: sprinkler irrigation of winter wheat in the North China
Plain. Agric. Water Manage. 159, 197208.
Wesseling, J., Elbers, J., Kabat, P., Van den Broek, B., 1991. SWATRE: Instructions for
Input. Internal Note. Winand Staring Centre, Wageningen, The Netherlands.
Xue, Q., Zhu, Z., Musick, J.T., Stewart, B.A., Dusek, D.A., 2003. Root growth and
water uptake in winter wheat under decit irrigation. Plant Soil 257, 151161.
Yang, G.-y., Luo, Y.-p., Li, B.-g., Liu, X.-y., 2006. The response of winter wheat root to
the period and the after-effect of soil water stress. Agric. Sci. China 5, 284290.
Yao, S., in Chinese with English abstract 2005. Effect of Sprinkler Irrigation on
Water Physiology, Photosynthesis, Crop Growth and Water Use Efciency of
Winter Wheat. Ph. D Thesis of Graduate University of Chinese Academy of
Sciences. University of Chinese Academy of Sciences, Beijing, China.
Yu, L.-p., Huang, G.-h., Liu, H.-j., Wang, X.-p., Wang, M.-q., 2009. Experimental
investigation of soil evaporation and evapotranspiration of winter wheat
under sprinkler irrigation. Agric. Sci. China 8, 13601368.
Zhang, H., Wang, X., You, M., Liu, C., 1999. Water-yield relations and water-use
efciency of winter wheat in the North China Plain. Irrig. Sci. 19, 3745.
Zhang, X., Pei, D., Li, Z., Li, J., Wang, Y., 2002. Management of Supplemental
Irrigation of Winter Wheat for Maximum Prot, Decit Irrigation Practices.
Food and Agriculture Organization, Rome, pp. 5765.
Zhang, X., Pei, D., Chen, S., 2004. Root growth and soil water utilization of winter
wheat in the North China Plain. Hydrol. Processes 18, 22752287.
Zhang, X., Chen, S., Sun, H., Wang, Y., Shao, L., 2009. Root size: distribution and soil
water depletion as affected by cultivars and environmental factors. Field Crops
Res. 114, 7583.
Zuo, Q., Jie, F., Zhang, R., Meng, L., 2004. A generalized function of wheats root
length density distributions. Vadose Zone J. 3, 271277.
van Genuchten, M.T., 1980. A closed-form equation for predicting the hydraulic
conductivity of unsaturated soils1. Soil Sci. Soc. Am. J. 44, 892898.