close

Вход

Забыли?

вход по аккаунту

?

s10333-017-0622-y

код для вставкиСкачать
Paddy Water Environ
DOI 10.1007/s10333-017-0622-y
ARTICLE
Enhancing field scale water productivity for several rice cultivars
under limited water supply
Roza Jonubi1 · Vahid Rezaverdinejad1 · Hamidreza Salemi2 Received: 5 June 2015 / Revised: 17 February 2017 / Accepted: 11 October 2017
© The International Society of Paddy and Water Environment Engineering and Springer Japan KK 2017
Abstract Rice production is one of the largest consumer of
water in agriculture. In general, the irrigation water productivity ­(WPI) is low in paddy fields. In order to improve ­WPI,
a field experiment was conducted in central of Iran during
2009–2010. The experiment was consisted of three irrigation managements and eight advanced rice cultivars (Gerdeh
(V1), Zayande-roud (V2), Sazandegi (V3), Hasani (V4), 67–97
(V5), 67–113 (V6), 67–47 (V7), and 67–72 (V8)) in a split
plot design with three replications. The irrigation treatments
were I1 and I2: permanent flooding under 3.5 and 2.2 cm
water depth, respectively, and I3: 0–1.5 cm alternative wetting and drying. To explore deficit irrigation for improved
­WPI, SWAP model was calibrated using intensive measured
data for the foregoing years. The average normalized rootmean-square deviation of yield during calibration was 0.03%
and during validation was 4.94% indicating acceptable calibration and validation of the model. ­WPI for all cultivars
were enhanced up to 61% by applying 50% deficit irrigation.
On this irrigation regime, V2 and V6 provided the highest
­WPI (0.84 and 0.79 kg m−3, respectively) whereas V4 and
V8 yielded the lowest (0.50 and 0.57 kg m−3, respectively).
The results indicated that rice cultivars (V2 and V6) are the
* Vahid Rezaverdinejad
v.verdinejad@urmia.ac.ir
Roza Jonubi
rjonubi@yahoo.com
Hamidreza Salemi
hr_salemiuk@yahoo.com
1
Department of Water Engineering, Urmia University, Sero
Road, Nazlou Campus, Urmia, P. O. Box: 165, Iran
Agricultural Engineering Research Department, Isfahan
Agricultural and Natural Resources Research and Education
Center, AREEO, Isfahan, Iran
2
best option with the highest ­WPI in the irrigation district and
calibrated model was able to effectively simulate the crop
growth under water deficit conditions.
Keywords Crop growth model · Deficit irrigation · Rice ·
SWAP · Water productivity
Introduction
Water demand for agricultural, industrial, and domestic is
growing rapidly. In other words, less water is available for
agriculture, in particular for high water demanding crops like
rice. However, more rice is needed to feed the growing population. Therefore, minimizing water loss and enhancing crop
yield is crucial to agricultural water policy makers. In agriculture, it is aimed to produce higher yields with less water
consumption. Water is an expensive economic good, and in
many cases, in addition, we have to pay enormous environmental costs also (Loeve et al. 2003). Water productivity is
a useful indicator for quantifying the impact of decisions on
irrigation schedules with regard to water management. The
water costs and scarcity brings in mind to think about virtual
water through the trade of it at the international and intranational levels (Liu et al. 2007). However, under current
water scarcity conditions, the limited available water should
be used more efficiently in any agricultural system (Bessembinder et al. 2005). The efficiency of water consumed by the
crops can be expressed as water productivity ­(WPT) and it
is defined as the production per plant transpiration amount.
Generally, it is difficult to distinguish plant transpiration (T)
from soil evaporation (E) at field scale. Hence, instead of
­WPT, it would be more convenient to use ­WPET which is the
production per unit evapotranspiration (ET). Total dry matter yield (Y) may also be translated into marketable yield,
13
Vol.:(0123456789)
i.e. Ym. If irrigation water depth (I) and precipitation (P)
water is considered as ‘crop water use’ then ­WPI + P may be
used. In arid regions with very low precipitations, W
­ PI + P
may be reduced to ­WPI (Vazifedoust et al. 2008). Irrigation
water productivity ­(WPI), reflects the relationship between
irrigating input and output water, and could be used as a
useful indicator for irrigation and crop plant management
level. Improvement in W
­ PI can reflect the comprehensive
improvement in crop production and irrigation water use
efficiency. Therefore, quantitative information on water
productivity indicators is necessary for planning efficient
irrigation water management under water-scarce conditions. In order to specify the most effective farm strategy
for achieving ‘more crop per drop’, we need to understand
the interactions between soil, atmosphere, crop and water.
Taking into account the spatial variability of soil and land
use properties as well as crop growth development, simulation of the water balance components will certainly help
to improve water productivity under water shortage conditions (Vazifedoust et al. 2008). Soil, Water, Atmosphere and
Plant inter-relationships (SWAP) model has been developed
to simulate crop growth and soil water processes. Scenario
analysis of the model allows to explore more efficient ways
of crop and water management under different hydrological
conditions. However, agro-hydrological models use a large
number of input parameters which may cause low performance and large uncertainty in simulation results such as
actual evapotranspiration, deep percolation and dry matter
yield (Singh et al. 2006).
Water productivity, a concept expressing the value or benefit derived from the use of water, includes various aspects
of water management and is very relevant for arid and semiarid regions (Droogers and Bastiaanssen 2002; Molden and
Sakthivadivel 1999). Researchers have been working on
WP values ­(WPT, ­WPET, percolation plus evapotranspiration ­(WPET + Q), ­WPI + P and W
­ PI) for different crops under
different conditions. In some studies, WP values were computed using the SWAP simulated water balance components,
i.e. transpiration, evapotranspiration, irrigation, and the marketable yield (Singh et al. 2006; Vazifedoust et al. 2008). In
addition to SWAP model, SWAP–WOFOST (Govindarajan
et al. 2008), coupling of SWAP-EPIC and ArcInfo geographic information system (Jiang et al. 2015) and CropSyst
model (Zare et al. 2014) were used to determine the water
productivity values. Distributed hydrological models like
SWAT has also been widely used at watershed and basin
scale (Ahmadzadeh et al. 2015).
There are a few previous studies of spatial simulation
modeling, which are related to different issues, e.g., agrohydrological process, irrigation water optimization, crop
water productivity analysis, distributed ecosystem simulation (Ines et al. 2006; Noory et al. 2011; Singh et al. 2006;
Van der Knijff et al. 2010). However, field experiments are
13
Paddy Water Environ
usually site-specific, expensive, time consuming and several
components of the water and salt balances can not readily
be measured. An alternative is to combine field experiments
with modeling using models such as SWAP to generalize
field specific observations in a rapid, flexible and inexpensive way (Van Dam et al. 2008). This approach was used to
evaluate actual evapotranspiration, crop yield and WP values
by Singh et al. (2006). Also, Vazifedoust et al. (2008) calibrated and validated SWAP model for eight selected farmer
fields (wheat, fodder maize, sunflower and sugar beet), then
the ­WPI values for the main crops were computed using the
SWAP simulated water balance components and deficit irrigation scheduling and reduction of the cultivated area were
studied. Jiang et al. (2015) considered combined effects of
weather, crop, soil and irrigation factors for the main crops
(grain corn, forage corn, wheat and vegetables) on water
productivity by close coupling of an agro-hydrological
model (SWAP-EPIC) and ArcInfo geographic information
system. In another study, Jiang et al. (2016) considered six
irrigation treatments including three levels of irrigation
water amounts: 100, 67 and 50% of ET (evapotranspiration,
510 mm) and two levels of irrigation water salinity: 0.65 and
3.2 dS m−1 for spring maize grown on saline soil.
For rice crop, the effect of irrigation regime options on
water productivity was evaluated by CropSyst model. The
treatments were irrigation regimes including continuous
flooding irrigation and 5-day irrigation intervals. Based
on result, to assess irrigation regimes, the most suitable
treatments regarding water productivity and environmental aspects were 5-day irrigation regime (Zare et al. 2014).
In another study, a profound water productivity analysis
was carried out at experimental field for rice crop under
different water regimes such as flooded (FL), alternative
wet and dry (AWD) and saturated soil culture (SSC). The
water productivity values were determined using SWAP and
SWAP–WOFOST. The highest value of water productivity
was observed from the flooded treatment and lowest value
from the saturated soil culture in W
­ PT and W
­ PET (Govindarajan et al. 2008). Previous researches did not discuss the
impact of deficit irrigation practices on WP values for different advanced rice cultivars. Therefore, in this study, we try to
utilize the SWAP model to comprehensively evaluate effect
of deficit irrigation on WP values (­ WPT, ­WPET and W
­ PI) for
rice crop and finding the appropriate cultivars.
In this study, we presented a methodology for evaluation
of on-farm strategies under water limited conditions, and
water productivity analysis of irrigated crops using measured and a simulated data. A field experiment was conducted
in a research station in the center of Iran during 2009–2010.
In this study, SWAP model was used for simulation. The
sensitivity of model input parameters were investigated
using the Lane method (Lane and Ferreira 1980). The SWAP
model was calibrated and validated using the observations
Paddy Water Environ
for the eight advanced rice cultivars. Finally, the water productivity and its variation for irrigated rice cultivars under
water-limited conditions was simulated and analyzed. The
analysis of variance (ANOVA) was used for analyze the
effect of different irrigation water depths on the rice production quantitative indicators. The ultimate goals of this
research are finding the appropriate cultivars in the irrigation district and promoting water management techniques to
enhance water productivity for irrigated rice in water scarce
areas with minimum adverse impacts on the crop growth and
production potential.
Materials and methods
Study area
This study was carried out at Shahid Fozveh research station in west of Isfahan province, Iran, from 2009 to 2010
(Fig. 1). The area has an approximate altitude of 1612 m
above MSL and is located at latitude 32.6°N and longitude
51.6°E. The region was classified as arid and semi-arid with
an average rainfall of 125 mm. The annual minimum and
maximum temperatures were 14 and 36 °C, respectively.
Irrigation water was supplied from wells. The soil and water
analysis are shown in Tables 1 and 2. The electrical conductivity (EC) of the majority of soils under rice cultivation is
in excess of 6.2 dS m−1. The experiment was conducted in
a split plot design (RCBD) with three replications (plot size
of 15 m × 17.5 m). The main plots comprised three levels
of irrigation managements: 3.5 cm permanent flood (I 1),
2.2 cm permanent flood (I2) and 0–1.5 cm alternate wetting
and drying (AWD) (I3). The sub plots consisted of eight
advanced rice cultivars included: Gerdeh (V1), Zayande-roud
(V2), Sazandegi (V3), Hasani (V4), 67–97 (V5), 67–113 (V6),
67–47 (V7), and 67–72 (V8). The treatments were compared
for grain yield and water productivity. Figure 2 presented the
layout of the experiment in the study area.
SWAP model
SWAP model was applied to simulate all the terms of water
and salt balance, and to estimate relative yields (i.e. actual
over potential yield). SWAP is an integrated physically
based simulation model for water, solute and heat transport
in the saturated–unsaturated zone in relation to crop growth.
SWAP is based on the finite difference solution of Richards’
equation extended with a sink term to account for root water
uptake. The first version of SWAP model was written in
1978 (Feddes et al. 1978) and since then, the model was
Fig. 1 Location of the study
area at Lenjanat irrigation district, Zayande-roud basin
13
15
34.5
1.6
56
29
28
84
63.3
7.6 22
6.2
In this study, the parameters of SWAP can be categorized
into soil hydraulic parameters, crop, irrigation and weather
data which are described below.
49.2
28.8
22
continuously updated (Droogers et al. 2000). The version
used for this study is SWAP3.2 and is described by Van Dam
et al. (1997). The core part of the program is the vertical
flow of water in the saturated–unsaturated zone, which can
be described by the well-known Richards’ equation. The
upper boundary condition in SWAP is determined by the
potential evapotranspiration ­(ETp), I and P. Daily model
outputs include simulated actual evaporation (Ea), actual
transpiration (Ta), and flow across the bottom of soil profile
and moisture distribution in the soil profile.
The input data files of SWAP are divided over four different types: main input file, meteorological file, crop growth
file, drainage file. The main input file and the meteorological
data file are always required. Input files of crop growth and
drainage are optional. The main input file contains general
information with regard to the simulation, meteorology, crop
rotation scheme, irrigation, soil water flow, lateral drainage,
bottom boundary, heat flow and solute transport. The meteorological file contains radiation, temperature, vapor pressure, wind speed, rainfall and/or reference evapotranspiration and rainfall intensities. Crop growth file are divided over
two different file types: file with detailed crop growth and
file with simple crop growth. In the present study, detailed
crop growth model was used. The detailed crop growth input
file is required to simulate crop development and biomass
assimilation. File with detailed crop growth contains two
sections: crop section and calculated irrigation section. Crop
section contains crop height, crop development, initial values, green surface area, assimilation, assimilates conversion
into biomass, maintenance respiration, dry matter partitioning, death rates, crop water use, salt stress, interception,
root growth and density distribution. Calculated irrigation
section contains irrigation time criteria and irrigation depth
criteria. The simple module describes crop development
independent of external stress factors. Its main function is
to provide proper upper boundary conditions for soil water
movement. The simple model is useful when crop growth
does not need to be simulated or when crop growth input
data are insufficient (Kroes et al. 2008). Drainage file and
simple crop growth file were not used in the present study.
Model parameterization
Clay
Mg2+ + Ca2+
­(meqL−1)
Texture Clay (%) Silt (%) Sand (%) EC (dS ­m−1) PH K (ppm) P (ppm) N (%) Na+
­(meqL−1)
Table 1 Physical and chemical analysis of soil samples
13
1.25
Paddy Water Environ
Cl− ­(meqL−1) HCO3−
FC ­(cm3 ­cm−3) PWP ­(cm3 ­cm3) ρb (g cm−3)
­(meqL−1)
Soil hydraulic properties
Soil hydraulic functions together with water retention and
hydraulic conductivity curves are required in order to simulate water flow through the soil profile. Soil hydraulic functions are not often available and, moreover, special apparatus
Paddy Water Environ
is required to determine these properties. Pedotransfer functions (PTF) can be used to derive these difficult-to-measure
soil hydraulic functions from easily obtainable data such as
soil texture and bulk density. Wösten et al. (1998) developed
a set of PTF using a soil database including data of 4030
horizons. These PTF were used to obtain the soil hydraulic
properties required as described according to Mualem–Van
Genuchten equations (Van Genuchten 1980). Variation in
soil properties was limited in the area and therefore, only
one soil type was considered. The parameters required to
describe soil hydraulic properties of the study area from
PTF were soil particles (clay, silt and sand), bulk density
(ρb), field capacity water content (θfc) permanent wilting
point water content (θpwp) and saturated water content (θsat).
The ρb b, θfc fc and θpwp pwp values of the soils were presented in Table 1. The soil hydraulic properties required as
described according to Mualem–Van Genuchten equations
are residual water content (res ), empirical shape factors
(
)
( and n), and the saturated hydraulic conductivity Ksat .
These parameters are presented in Table 3.
Crop and irrigation data
A summary of the SWAP input data with regard to crop
type and irrigation schedules is shown in Table 4. The maximum crop height was obtained from field observations. Leaf
area index (LAIEM) for the initial stage (emergence) was
selected from the literature (Allen et al. 1998) and maximum
rooting depth as a function of the crop development stages
was measured at the monitored fields. The applied irrigation
water depths were measured at the monitored fields. Precipitation interception coefficient and LAI as a function of crop
growth stage are used for calculation of the precipitation
interception by crop canopy using Braden (1985) formula.
Intercepted water is evaporated from the crop canopy and
Table 2 Chemical analysis of irrigation water
EC (dS ­m−1)
PH
CO3H− ­(meqL−1) Cl− ­(meqL−1)
SO42− ­(meqL−1)
Total onions
­(meqL−1)
Mg2+ + Ca2+
­(meqL−1)
Na+ ­(meqL−1)
Total cations
­(meqL−1)
3.9
7.4
2.4
8.6
35
17
18.5
35.5
24
Fig. 2 Layout of the experiment in the research field
Table 3 Soil texture and soil hydraulic parameters in the research field
Layer depth (m)
Clay (%)
Silt (%)
Sand (%)
θsat ­(cm3 ­cm−3)
θres ­(cm3 ­cm−3)
α ­(cm−1)
N (−)
Ksat (m day−1)
0–60
49.2
28.8
22
0.473
0.094
0.017
1.276
0.3
13
Paddy Water Environ
Table 4 Crop and irrigation data used in SWAP model
Parameter
Unit
Min
Max
First year (2009)
Second year (2010)
Determination
General
Simulation period
Crop
–
–
–
–
–
–
22 May–17 Oct
The eight advanced
rice cultivars
31 May–07 Oct
The eight advanced
rice cultivars
E
E
mm
mm
mm
–
–
–
–
–
–
1700
1078
705
1525
880
627
M
M
M
day
cm
cm
cm
cm
S m−1
dS m−1
% ­dS−1 m
o
C
o
C
kg ha−1
m2 m−2
m2 m−2 d
–
–
kg ha−1 h−1
kg kg−1
–
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
–
1000
1000
1000
1000
1000
20
40
10,000
10,000
10,000
10
1
2
2
10
1
118
115
105
95
50
70
3
12
740
745
138
0.09
0.007
0.4213
0.9785
0.46
0.415
118
110
100
90
50
70
3
12
724
705
138
0.09
0.007
0.4213
0.9785
0.46
0.415
E
M
M
M
M
L
L
L
M
M
D
L
D
D
D
D
D
Irrigation data
Irrigation water quantity in treatment I1
Irrigation water quantity in treatment I2
Irrigation water quantity in treatment I3
Crop parameters
Length of crop cycle
Maximum crop height in treatment I1
Maximum crop height in treatment I2
Maximum crop height in treatment I3
Maximum rooting depth
Minimum canopy resistance
ECMAX
ECSLOPE
TSUMEA
TSUMAM
TDWI
LAIEM
RGRLAI
KDIF
KDIR
EFF
CVO
D default, M measured, L literature and E local experience
transpiration component is treated as zero. Potential soil
evaporation rate is derived from Penman–Monteith equation by neglecting the aerodynamic term and assuming that
the net radiation inside the canopy decreases according to
an exponential function, and that the soil heat flux can be
(Van Dam et al. 1997; Vermaa et al. 2012) neglected. Plant
parameters as initial total crop dry weight (TDWI), maximum relative increase in LAI (RGRLAI), extinction coefficient for diffuse visible light (KDIF), extinction coefficient
for direct visible light (KDIR), light use efficiency for real
leaf (EFF) and efficiency of conversion into storage organs
(CVO) were considered in Table 4 by default. Temperature sum from emergence to anthesis (TSUMEA) and temperature sum from anthesis to maturity (TSUMAM) were
measured. ­ECe level at which salt stress starts (ECMAX),
decline of root water uptake above ECMAX (ECSLOPE)
and minimum canopy resistance were selected from the literature (Allen et al. 1998). There are three basic ways to
determine irrigation needs, first; hand test of the soil, second; using tensiometers option, third; using water budget
approach. Third method is based on climatic data and has
13
the following advantages for scheduling irrigation: (1) no
equipment requirements; (2) accuracy; (3) simplicity of use;
and (4) flexibility allowing easy adaptation for use in other
crops. In this study, third method was used to determining
irrigation needs. Whereas, climatic conditions were different
in the 2 years (2009 and 2010), the irrigation applied under
both years were different (Table 5).
Weather data
The meteorological data on a daily basis including daily
rainfall, maximum and minimum temperature, wind velocity, relative humidity, sunshine hours and solar radiation
were collected from Kabutar–Abad meteorological station
(Fig. 3).
Evapotranspiration
Evapotranspiration refers to both transpiration of the plants
and evaporation from the soil or of water intercepted by
vegetation or ponding on the soil surface. In SWAP, the
Paddy Water Environ
Table 5 Irrigation scheduling
for three treatments (I1, I2 and
I3) in 2009 and 2010
Year
2009
Treatment
I1
I2
I3
2010
I1
I2
I3
Parameter
Month
Initial date of each irrigation
Frequency (times a month)
Average depth of irrigation (mm)
Monthly depth of irrigation (mm)
Initial date of each irrigation
Frequency (times a month)
Average depth of irrigation (mm)
Month depth of irrigation (mm)
Initial date of each irrigation
Frequency (times a month)
Average depth of irrigation (mm)
Month depth of irrigation (mm)
Initial date of each irrigation
Frequency (times a month)
Average depth of irrigation (mm)
Monthly depth of irrigation (mm)
Initial date of each irrigation
Frequency (times a month)
Average depth of irrigation (mm)
Monthly depth of irrigation (mm)
Initial date of each irrigation
Frequency (times a month)
Average depth of irrigation (mm)
Monthly depth of irrigation (mm)
May
Jun
Jul
Aug
Sep
22
4
40
160
22
5
22
110
22
4
13.5
54
31
1
37
37
31
1
22
22
31
1
20
20
1
13
35
455
1
14
22
308
2
13
14
182
3
12
36.5
438
2
14
22
308
2
13
16.5
214
2
13
35
455
4
13
22
286
1
15
13.0
197
1
13
35
455
1
11
22
242
1
12
13.5
160
2
11
35
385
4
11
22
242
1
14
12.0
168
3
10
35
350
1
8
22
176
1
11
13.7
150.9
2
7
35
245
4
6
22
132
1
8
13.0
104
2
7
35
245
4
6
22
132
3
8
10.3
82.4
Fig. 3 Monthly meteorological data for 2009 and 2010
atmospheric conditions are assumed to be external conditions, which are representative for the area for which the
simulations are performed. Starting point in the calculations is the determination of the potential evapotranspiration of different uniform surfaces. The addition ‘potential’
refers to non-limiting water supply from the soil. The model
offers two methods to calculate this potential evapotranspiration: the Penman–Monteith method and the reference
evapotranspiration method. The SWAP model can separate
ET, to the E and T is at different stages of plant growth.
In SWAP the crop factors are used to convert the evapotranspiration rate of a reference crop, fully covering the soil
to the potential evapotranspiration rate of the actual crop,
fully covering the soil. This is different from programs
like CROPWAT, which use crop factors that depend on the
crop development stage and soil cover. Because the soil has
13
generally a dry top layer, soil evaporation is usually below
the potential evaporation rate. Hence, these crop factors
semi-empirically combine the effect of an incomplete soil
cover and reduced soil evaporation. Therefore, crop factors in SWAP can be larger than those in CROPWAT. In
this study, SWAP has built in alternatives for calculation of
actual crop evapotranspiration. Potential crop evapotranspiration is calculated by Penman–Monteith equation for a reference crop, and is separated into potential crop transpiration
and soil evaporation. Measured potential evapotranspiration
were calculated from basic meteorological data using the
Penman–Monteith equation and then the crop factors were
used to convert them to actual evapotranspiration.
Calibration and validation
The data collected in year 2009 were used to calibrate the
model while those of 2010 were used to validate the model.
Calibration was performed for dry matter production. During
model calibration, the simulated output of SWAP model was
compared with the observed data of dry matter production of
different cultivars. Each cultivar was defined to the model during calibration of plant parameters. The calibration was accomplished with minimal difference of the model simulated and the
observed results corroborated through statistical analysis. Simulation of dry matter production in SWAP was sensitive to plant
parameters and difference between rice cultivars were resulted
them, because plant parameters as TDWI, LAIEM, RGRLAI,
KDIF, KDIR, EFF and CVO were estimated from the calibration procedure. In the following, their sensitivity analyses were
investigated. Further, the calibrated SWAP model was validated
for different varieties using the experiment data of 2010. During
the model validation process, only the weather and irrigation
scheduling parameters acquired from the experiment during
2010 were used as inputs in the calibrated model.
Simulation scenarios
SWAP model was calibrated and validated for eight advanced
rice cultivars in order to specify maximum possible water productivity and explore on-farm strategies for higher economic
yields. As on-farm strategy, a number of deficit irrigation
schedules, were analyzed. All parameters in the model were
kept constant over all the farms, except the depth of irrigation.
The irrigation depth was also defined based on percent irrigation (20, 30, 40, 50, 60, 70, 80, and 90%). Water productivity
was calculated using the simulated water balance components
T, ET, I and the actual yield, Ym. Water productivity-water
consumption curves were presented for all the rice cultivars
and for different irrigation practices. The maximum possible
enhancement in water productivity indicators was derived
graphically from the water productivity curves.
13
Paddy Water Environ
Sensitivity analysis
Parameters sensitivity analysis is a prerequisite step in the
model-building process. The sensitivity analysis method
identifies parameters that do or not have a significant impact
on model simulation and is critical for reducing the number
of parameters required in model validation. In this study,
the sensitivity of crop input parameters were investigated
using the Lane method (Lane and Ferreira 1980).According
to Lane method, the model was run using the experiment
data of 2009 used for model calibration and the model output
results were compared with the observed data. Moreover, the
model was simulated repeatedly with change in the value
of one crop input parameter at a time by 25% increase and
a time by 50% increase and then by 25% decrease and 50%
decrease of its original value. Crop input parameters were
TDWI, LAIEM, RGRLAI, KDIF, KDIR, EFF and CVO.
This process was repeated for all input values and the results
were compared with the observed data (i.e. base value) of
yield and the absolute differences were determined using
Eq. (1):
|M − I |
| × 100
D = ||
|
| I |
(1)
where D is absolute difference between the output value and
the base value, I is base value and, M is output value. Further, the sensitivity of the model was determined under the
following conditions:
D = 0 Not sensitive
0 < D
< D ≤ 0 Low sensitive
10 < D ≤ 50 Medium sensitive
D > 50 High sensitive
Data collection
During 2009 and 2010, detailed water, soil and crop parameters were collected from farmers within the study area. Table 6
shows a summary of the data used for calibration and validation of SWAP model.
SWAP performance evaluation
The model performance was evaluated using six standard
statistical measures: root mean squared error (RMSE), the
normalized root-mean-square error (NRMSE), coefficient
of determination (R2), coefficient of residual mass (CRM),
Nash–Sutcliffe efficiency (NSE) and relative error (RE). These
statistical measures provide more information on the systematic and dynamic errors occurring in the model simulation.
Paddy Water Environ
RMSE shows how much the simulations under- or over-estimate the measurements. RMSE is computed as:
�
�2
∑N �
Yobs − Ysim
i=1
(2)
RMSE =
N
where N is number of observations, Yobs is measured value,
Ysim is simulated or predicted value.
NRMSE is calculated with the following equation:
�
�2
∑N �
Yobs − Ysim
1
i=1
(3)
× 100
NRMSE =
N
Ȳ obs
where Ȳ obs is the mean measured value. An NRMSE value
smaller than 10% is ideal for modeling. An NRMSE in the
range of 10 to 20%, or 20 to 30%, indicates an appropriate
or moderate condition in model predictions, respectively.
An NRMSE in excess of 30% indicates uncertainty of the
model. R2 is defined as the squared value of the coefficient
of correlation. It is calculated as (Krause et al. 2005):
2
⎛ ∑�
⎞
�
Yobs × Ysim − n × Ȳ obs × Ȳ sim ⎟
⎜
2
R = ⎜ ��
��∑ 2
�⎟
∑ 2
2
2
⎜
⎟
Yobs − nȲ obs
Ysim − nȲ sim
⎝
⎠
(4)
where is Ȳ sim is the mean simulated or predicted values.
The range of R2 lies between 0 and 1, and describes how
much of the observed dispersion is explained by the prediction.
A value of zero means no correlation at all; whereas a value
of 1 means that the dispersion of the prediction is equal to that
of the observation.
CRM is a measure of the tendency of the model to overestimate or under-estimate the measurements. Positive values
for CRM indicate that the model underestimates the measurements and negative values indicate a tendency of the model
for overestimation. CRM is computed as:
∑
CRM =
∑
Yobs − Ysim
∑
Yobs
(5)
NSE is a normalized statistic that determines the relative
magnitude of the residual variance compared to the measured
data variance. NSE indicates how well the plot of observed
versus simulated data fits the 1:1 line. NSE is computed as
shown in Eq. (6):
∑�
�2
Yobs − Ysim
NSE = 1 − ∑ �
�2
Yobs − Ȳ obs
(6)
Table 6 Overview of the data used in calibration and validation of SWAP model
Data
Collection method
Frequency
Meteorological data
Soil properties
Texture
Bulk density
Saturated hydraulic conductivity
Saturated percentage
Organic Carbon
pH
Soil moisture
Irrigation regime
Discharge of irrigation source, i.e. canal water or tube-well water
Duration of irrigation
Irrigation depth
Meteorology station
Daily
USDA method
Core method
Constant water head method
Saturated paste method
Wet digestion method (Walkley and Black 1934)
In soil–water suspension of 1:2
Gravimetric method
Once
Once
Once
Once
Before, Sowing
Before, Sowing
Weekly
Current meter/co-ordinate/volumetric method
Field observation
By multiplying the discharge and duration of irrigation and then divided by field area
Field observation
Conductivity meter
Once
Once
After each irrigation
After each irrigation
Once
Field observation
Once
Field observation
Field observation
Field observation
Field observation
Field observation
4–5 times
4–5 times
4–5 times
4–5 times
Once
Irrigation date
Irrigation quality
Crop growth parameters
Crop development stages, i.e. emergence, panicle initiation, anthesis, maturity and harvest
Plant density and tillers
Plant height
Dry matter partitioning
Rooting depth
Crop yield
13
Paddy Water Environ
NSE ranges between −∞ and 1.0 (1 inclusive), with
NSE = 1 being the optimal value. Values between 0.0 and
1.0 are generally viewed as acceptable levels of performance,
whereas values < 0.0 indicate that the mean observed value
is a better predictor than the simulated value, which indicates
unacceptable performance (Nash and Sutcliffe 1970). Finally,
RE is computed as:
|Yobs − Ysim |
| × 100
RE = |
Yobs
(7)
Percent deviation between calibrated crop parameters
and their default values were determined using the following equation:
Percent deviation =
(calibrated − default)
× 100
default
(8)
Results and discussion
SWAP model calibration and validation results
The observed sensitivities of different crop input parameters are presented in Table 7. With respect to Table 7, EFF
parameter has high sensitivity, RGRLAI and CVO parameters have moderate sensitivity. Also, TDWI, KDIF, LAIEM
and KDIR parameters have low sensitivity.
Plant parameters including TDWI, LAIEM, RGRLAI,
KDIF, KDIR, EFF and CVO were determined from the calibration procedure. The default and calibrated crop parameters for the eight advanced rice cultivars were presented in
Table 8. Also, percent deviation for each cultivar calibrated
crop parameters from default were given in Table 8. The
average percent deviation for TDWI, LAIEM, RGRLAI,
KDIF, KDIR, EFF and CVO parameters were 5.35, 0, 0,
2.85, 0.1, 1.06 and 30.42, respectively. Maximum average percent deviation parameter was CVO and the minimum parameters were LAIEM and RGRLAI parameters.
The most difference on eight advanced rice cultivars were
between CVO and TDWI parameters.
The model performance after the calibration and validation processes was evaluated by comparing the measured
with the simulated results. The agreement between the
simulated and the observed values acquired from field
experiments was statistically evaluated by estimating the
RMSE, NRMSE, RE, R2, CRM and NSE. All the statistical
parameters for each rice cultivar were presented in Table 9.
The mean and maximum NRMSE for all the cultivars were
0.03 and 0.06%, respectively. Since both of them were
less than 10%, they are ideal for modeling. The mean and
maximum RE values were 0.03 and 0.04%, respectively,
indicating a fair accuracy in the yield prediction. The
13
average and the maximum RMSE were 2.5 and 4.76 kg,
respectively, which are negligible. The NSE, CRM and
R2. values were 1, 0 and 1, respectively. Therefore, the
observed and predicted dry matter productions were in
close agreement (Figs. 4 and 5). Validation results were
shown in Table 9. Although RMSE values on calibration
were much less than validation, but they were acceptable. Also, RMSE values were similar with the results of
other studies that in the following, they were presented.
The NRMSE is dimensionless and performs better than
RMSE for model performance evaluation analyzes. Since
the average NRMSE for all the cultivars was 4.94% and
all NRMSE values, except for V4 (10.76%) were less than
10%, therefore, it is ideal for modeling. The other statistical parameters indicated an acceptable validation of the
model. The most and the least accurate predictions were
in cultivars V7 and V4, respectively.
Amiri et al. (2014) evaluated CERES-Rice, AquaCrop
and ORYZA2000 models performance in simulation of
grain yield of rice in response to different irrigation intervals and nitrogen levels. These models were calibrated
and validated by using 3 years (2005 to 2007) field experiments. The NRMSE were determined 16, 15 and 23%
in CERES-Rice, AquaCrop and ORYZA2000 models,
respectively. In another study, management practices for
increasing rice production determined in Laos by using
the CSM-CERES-Rice model. The average RMSE and
NRMSE values were 145 kg ha −1 and 4%, respectively
(Vilayvong et al. 2015). Confalonieri et al. (2009) have
used the WARM model for the simulation of rice growth
under flooded and un-flooded conditions in China and
Italy. The average NRMSE, CRM and R2 of aboveground
biomass were 21.9%, 0.06 and 0.94, respectively for the
calibration, and 23.9%, 0.13 and 0.96, respectively for
validation.
Evapotranspiration
The measured and simulated evapotranspiration for I 1
treatment were shown in Fig. 6. The difference of evapotranspiration between measured and simulated ET using
SWAP model were less than 9 and 5% in calibration step
Table 7 Sensitivity of SWAP model to different crop input parameters in simulating yield of rice
High sensitivity
Moderate sensitivity
Low sensitivity
EFF
RGRLAI
CVO
TDWI
KDIF
LAIEM
KDIR
Paddy Water Environ
Table 8 The default and calibrated crop parameters of the crop growth for the eight advanced rice cultivars
Cultivar
Parameter
Default
V1
Deviation (%)
V2
Deviation (%)
V3
Deviation (%)
V4
Deviation (%)
V5
Deviation (%)
V6
Deviation (%)
V7
Deviation (%)
V8
Deviation (%)
Average deviation (%)
TSUMEA
(oC)
TSUMAM
(oC)
TDWI
(kg ha−1)
LAIEM
­(m2 ­m−2)
RGRLAI
KDIF
­(m2 ­m−2 d) (−)
KDIR (−) EFF
CVO (kg kg−1)
(kg ha−1 h−1)
–
740
–
740
–
740
–
740
–
740
–
740
–
740
–
740
–
–
–
745
–
745
–
745
–
745
–
745
–
745
–
745
–
745
–
–
138
123.83
10.27
127.57
7.56
130.83
5.20
126
8.70
132
4.35
138
0.00
128.67
6.76
138
0.00
5.35
0.09
0.09
0.00
0.09
0.00
0.09
0.00
0.09
0.00
0.09
0.00
0.09
0.00
0.09
0.00
0.09
0.00
0.00
0.007
0.007
0.00
0.007
0.00
0.007
0.00
0.007
0.00
0.007
0.00
0.007
0.00
0.007
0.00
0.007
0.00
0.00
0.9785
0.982
0.36
0.979
0.05
0.979
0.05
0.979
0.05
0.979
0.05
0.979
0.05
0.979
0.05
0.98
0.15
0.10
0.4213
0.431
2.30
0.43
2.07
0.433
2.78
0.411
2.44
0.437
3.73
0.423
0.40
0.444
5.39
0.437
3.73
2.85
0.46
0.462
0.43
0.46
0.00
0.463
0.65
0.45
2.17
0.468
1.74
0.453
1.52
0.461
0.22
0.468
1.74
1.06
0.415
0.605
45.78
0.612
47.47
0.6
44.58
0.407
1.93
0.53
27.71
0.61
46.99
0.525
26.51
0.425
2.41
30.42
Table 9 Statistical parameters for calibration and validation steps of SWAP model
Cultivar
V1
V2
V3
V4
V5
V6
V7
V8
Average
Calibration
Validation
CRM (−) NSE (−) NRMSE (%) RE (%) RMSE (kg)
R
CRM (−) NSE (−) NRMSE (%) RE (%) RMSE (kg) R2
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
− 0.05
− 0.04
− 0.06
− 0.1
0.02
− 0.02
0.01
0.01
− 0.03
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.02
0.06
0.02
0.05
0.03
0.03
0.04
0.03
0.03
0.02
0.04
0.01
0.03
0.03
0.02
0.03
0.02
0.03
1.63
4.76
0.82
3.51
2.16
2.38
2.89
1.83
2.5
(2009 year) and validation step (2010 year), respectively,
which were acceptable.
Analysis of variance (ANOVA)
The analysis of variance (ANOVA) was used for rice production quantitative indicators including days after planting
to 50% maturity, days after planting to full maturity, root
length, root dry weight, crop height, spike length, 1000grain weight, grain yield, tiller number, rice grain width,
2
− 0.13
− 0.61
− 2.57
− 0.4
0.56
0.49
0.69
0.55
0.04
5.39
4.53
6.52
10.73
4.08
6.1
3.68
2.91
4.94
4.6
3.78
6.38
10.06
2.58
6.12
3
2.82
4.23
390.44
364.02
471.29
483.02
301.84
467.72
261.09
165.55
318.81
0.68
0.59
0.85
0.79
0.76
0.70
0.86
0.63
0.84
rice grain length, number of grains per spike and water use
efficiency (WUE) (Table 10). The effects of irrigation water
depth on the days after planting to full maturity at 5% level
of probability for the eight advanced rice cultivars was significant and in other indicators at 1% level of probability
were significant. Interaction between irrigation and cultivars
on the WUE, root dry weight, spike length and rice grain
width were significant at 1% level of probability, but on the
other indicators were not significant.
13
Fig. 4 Simulated and observed dry matter yields for the eight rice cultivars
Fig. 5 Simulated and observed dry matter yields in a calibration step, b validation step
Fig. 6 Observed and simulated cumulative ET for I1 treatment. a Calibration step, b validation step
13
Paddy Water Environ
Paddy Water Environ
Water productivity (WP)
Water productivities resulted from the irrigation treatments
The irrigation depth of I1, I2 and I3 treatments were 1700,
1078 and 705 mm, respectively. Grain yield for the aforementioned treatments are 7.65, 7.50 and 6.59 tonnes h­ a−1,
respectively. The WP resulted from the irrigation treatments
were calculated using simulated water balance components
T, ET and I for crop productions. The grain yield in I2 and
I3 treatments were less than I1 treatment, but WP in these
treatments were higher than I1 treatment (Table 11).
The V2 and V6 cultivars provided the highest and V4 and
V8 yielded the lowest water productivity irrigation district.
Given the irrigation water salinity of 4 dS m−1, it can be
concluded that V2 and V6 cultivars have relatively high tolerance to salinity. A comparison between the values of ­WPET
and ­WPT revealed that soil evaporation decreased W
­ PI in the
range of 23 to 29%. Furthermore, V2 cultivar provided the
highest, and V4 and V8 cultivars yielded the lowest ­WPET.
Differences in WP within different crops could be attributed to the differences in soil chemical composition, harvest index and transpiration demands (Singh et al. 2006). By
comparing the values of ­WPET and W
­ PI observed that W
­ PI
was smaller than W
­ PET because of deep percolation and low
irrigation efficiency.
Increasing the water productivity of irrigated crops
Deficit irrigation is an optimization strategy in which irrigation is applied during drought-sensitive growth stages of
a crop. Outside these periods, irrigation is limited or even
unnecessary if rainfall provides a minimum supply of water.
Water restriction is limited to drought-tolerant phenological stages, often the vegetative stages and the late ripening
period. Total irrigation application is therefore not proportional to irrigation requirements throughout the crop cycle.
While this inevitably results in plant drought stress and consequently in production loss, deficit irrigation maximizes
irrigation water productivity, which is the main limiting factor (English 1990). In paddy field, soil should be saturated,
so for applying deficit irrigation, irrigation water depth was
reduced during growth period. There are different ways to
manage deficit irrigation. An irrigator may reduce the irrigation depth, refill only part of the root zone, reduce the irrigation frequency by increasing the interval between successive
irrigations, and refill a fraction of the soil water capacity of
the root zone, wetting furrows alternately or placing them
further apart. The association of high WP values with high
yields has important implications on the crop management
for achieving efficient use of water resources in water-scarce
­ PET versus IR
areas. Irrigation water productivity (­ WPI) or W
or ET curves clarify the importance of attaining relatively
higher yields with higher water productivity. Attaining
higher yields with increased WP is economical only when an
increase in crop yield is not offset by an increase in the costs
of other inputs. Policies for maximizing the yield should be
considered carefully before being applied in water-scarce
conditions. Guidelines for recommending irrigation schedules under normal water availability may need to be revised
when applied in water-scarce areas. WP was calculated using
the simulated water balance components T, ET, I and the
actual marketable yield, Ym. Figure 7 shows the relationship between WP indicators in terms of Ym T−1, Ym ET−1
and Ym I−1. Three fluxes comprising evaporation, percolation
and soil moisture storage in the soil profile are considered
as losses and are the main causes for reduction of WP from
­WPT to W
­ PET and W
­ PET to W
­ PI (Vazifedoust et al. 2008). Of
course in paddy fields, soil is saturated and moisture content
which stored in the soil profile isn’t considered as loss. The
relationship between ­WPET and the amount of water used for
ET demonstrates that an increase in water consumption is
accompanied by a proportionate increase in W
­ PET. However,
the relationship between ­WPET and ET depends on several
factors such as stability of harvest index (HI), plant density
per unit area, nutrient supply and fertilizers, and crop cultivar (Keller and Seckler 2005). Also, when water supply
is not sufficient to keep the actual transpiration of a crop
equal to potential transpiration, the stomata of the leaves will
close partially causing a decline in yield, as a result, W
­ PI
does change with amount irrigation. The most outstanding
conclusion is that crop ­WPI can be increased significantly if
irrigation is reduced and crop water deficit is intentionally
induced.
As shown in Fig. 7 and Table 12, application of deficit irrigation led to an increase in water productivity.
Based on the simulation outputs, W
­ PI for all cultivars will
increase up to 61% by applying 50 percent deficit irrigation. According to Table 12, it can be seen that by applying 50% deficit irrigation, V2 and V6 provided the highest
water productivity (0.84 and 0.79, respectively) whereas
V4 and V8 yielded the lowest (0.50 and 0.57, respectively).
In spite of improvements in ­WPI, justification for shifting from traditional to modern irrigation practices is not
straightforward. A main reason is that maximum water productivity will often not coincide with farmers’ interests,
maximize their yields or economic profitability. Maximizing ­WPI requires a shift in irrigation science, irrigation
water management and financial investment to move away
from ‘maximum irrigation-maximum yield’ strategies to
‘less irrigation maximum WP’ policies which is often not
a common interest between farmers and water managers
(Zwart and Bastiaanssen 2004).
13
13
27.2n.s
135.8n.s
10.2
14
–
9.6
9.7*
29.6**
28.3
789.1**
2352**
431.4
2239**
833.6**
486.6**
391.4
1028.5**
1556**
Root dry
weight
(gr)
n.s Not significant, ** significant at p < 0.01, * significant at p < 0.05
DAP Days after planting
55.2*
195.4n.s
7
6.2
53
54.9*
27.3n.s
4.266
377**
121.7n.s
8
7
14
42,569
15.8**
9.8n.s
3895**
730,852** 70.8**
543
52
48.1
3.5n.s
27.9**
368n.s
232.7**
7.203n.s 3.9n.s
1
4
2
2
Year
Repeat (year)
Irrigation
Irrigation × Year
Error
Cultivar
Irrigation × Cultivar
Cultivar × Year
Year × Cultivar × Irrigation
Coefficient
of variation
(CV %)
Root
length
(cm)
Degrees
of freedom
Source of
changes
DAP to
DAP to
50% matu- full maturity (day) rity (day)
5.6
65.7*
689**
35,065
584**
43.4n.s
2678**
1034
361**
251**
Crop
height
(mm)
6.2
0.99n.s
1.7n.s
42,376
34.9**
4.2**
16**
1.22
17.2**
13.3**
Spike
length
(mm)
4
3.1*
3.8**
42,553
165**
1.11n.s
28.3**
42,435
7.8n.s
2.1n.s
9.4
0.35n.s
4**
4
24.9**
0.52n.s
10.4**
42,463
4.4n.s
2.1*
12.7
18.7*
23.7*
42,471
31**
14.6n.s
29.3n.s
28,946
12.1n.s
0.46n.s
1000 grain Grain yield Tiller
(kg ha−1)
weight
number
(gr)
4
0.04**
0.007n.s
0.026
0.19**
0.042**
0.027n.s
0.019
0.004n.s
0.028n.s
2.5
0.049n.s
0.024n.s
0.068
3.5**
0.06n.s
0.29*
0.19
0.034n.s
0.0007n.s
13.5
118.2n.s
281.2n.s
462
4911**
82.8n.s
1914**
190
271.2n.s
49n.s
Rice grain Rice grain Number
of grains
length
width
per spike
(mm)
(mm)
Table 10 Compound analysis of variance (ANOVA) for study the effect of different irrigation depths on the rice production quantitative indicators
1.95
0.000058n.s
0.000067n.s
0.000058
0.18**
0.0059**
0.013**
0.00,069
0.145**
0.000096**
Water use
efficiency
(kg m−3)
Paddy Water Environ
Paddy Water Environ
Table 11 Water productivity indicators ­WPT, ­WPET and ­WPI (kg m−3) in irrigation treatments and rice cultivars
Cultivar treatment
V1
V2
V3
V4
I1
I2
I3
I1
I2
I3
I1
I2
I3
I1
I2
I3
WPT
WPET
WPI
1.31
0.99
0.47
1.30
0.98
0.75
1.24
0.89
1.00
1.42
1.08
0.52
1.42
1.08
0.83
1.36
0.99
1.11
1.31
0.99
0.48
1.31
0.99
0.76
1.26
0.92
1.04
0.87
0.65
0.31
0.87
0.65
0.50
0.82
0.58
0.65
Cultivar treatment
V5
WPT
WPET
WPI
V6
V7
V8
I1
I2
I3
I1
I2
I3
I1
I2
I3
I1
I2
I3
1.21
0.93
0.44
1.21
0.93
0.72
1.16
0.86
0.97
1.35
1.02
0.49
1.35
1.01
0.77
1.29
0.93
1.05
1.18
0.90
0.43
1.18
0.90
0.69
1.13
0.83
0.94
0.96
0.73
0.35
0.96
0.73
0.56
1.04
0.76
0.85
Fig. 7 Relationship between WP indicators (kg m−3) in terms of Ym T−1, Ym ET−1 and Ym I−1, and irrigation water depth
Conclusions
In this study, the SWAP model was calibrated and validated
using measured data at eight selected farmer’s fields (V1 to
V8) in Shahid Fozve research station, Isfahan. The PTF were
used to determine indirectly the remaining uncertain soil
hydraulic parameters (res, , Ksat and n). Also, some crop
parameters were determined during calibration for all the
cultivars. The sensitivity of different crop input parameters
were observed on the basis of Lane method (Lane and Ferreira 1980). Based on the results, EFF parameter has high
sensitivity, RGRLAI and CVO parameters have moderate
sensitivity and TDWI, KDIF, LAIEM and KDIR parameters have low sensitivity. For calibration and validation,
the simulated output of SWAP model was compared with
the observed data of dry matter production for different cultivars. The average NRMSE for all the cultivars were 0.03
and 4.94% for calibration and validation steps, respectively.
Since both of them are less than 10%, they are ideal for
modeling. The average RE values were 0.03 and 4.23% for
calibration and validation respectively and the NSE, CRM
and R2 values were 1, 0 and 1 for calibration and 0.04,
− 0.03 and 0.84 for validation, respectively. Based on analysis of variance (ANOVA), effect of the eight advanced rice
cultivars on the rice production quantitative indicators were
significant at 1% level of probability. The effects of different
irrigation depths on the WUE, root length, root dry weight,
crop height and spike length were significant at 1% level of
probability. The effect of interaction between irrigation and
cultivars on the WUE, root dry weight, spike length and
rice grain width were significant at 1% level of probability.
To explore on-farm strategies, a number of deficit irrigation
schedules were analyzed using the calibrated SWAP models.
The irrigation depth was defined based on percent irrigation
(20, 30, 40, 50, 60, 70, 80, and 90%). The WP values of the
main crops were computed in different forms of W
­ PT, ­WPET
and ­WPI. By comparing the values ­WPI for full irrigation
treatment observed that V2 and V6 cultivars provided the
highest water productivity and V4 and V8 cultivars provided
the least water productivity in the irrigation district. According to source of irrigation water with salinity of 4 dS m−1,
it can be concluded that V2 and V6 cultivars have high tolerance to salinity. Based on the results, ­WPI for all cultivars
will increase up to 61% by applying 50% deficit irrigation
scheduling. Decreasing irrigation amount led to decreased
yield, but the water required obtaining maximum ­WPI was
much lower than that required for obtaining maximum yield.
Moreover, higher water efficiency would be achieved with
13
Table 12 Water productivity
indicators ­WPT, ­WPET and
­WPI for various irrigation
schedules (90, 70, 50 and 30%
of irrigation depth)
Paddy Water Environ
Cultivar
Irrigation (%)
Crop yield
(tonnes
­ha−1)
Water
consumption
(mm)
WPT (kg m−3)
WPET (kg m−3)
WPI (kg m−3)
V1
90
70
50
30
90
70
50
30
90
70
50
30
90
70
50
30
90
70
50
30
90
70
50
30
90
70
50
30
90
70
50
30
8.05
7.98
6.47
2.91
8.75
8.67
7.12
3.30
8.05
7.98
6.47
2.92
5.25
5.20
4.24
1.87
7.52
7.46
6.07
2.76
8.22
8.15
6.66
3.03
7.38
7.28
5.93
2.70
5.95
5.89
4.79
2.17
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1525.50
1186.50
847.50
508.50
1.31
1.31
1.19
0.79
1.42
1.42
1.30
0.89
1.31
1.30
1.19
0.79
0.87
0.87
0.79
0.52
1.21
1.20
1.10
0.73
1.35
1.34
1.23
0.83
1.18
1.17
1.07
0.72
0.96
0.96
0.87
0.58
0.99
0.99
0.87
0.50
1.08
1.07
0.96
0.57
0.99
0.99
0.87
0.50
0.65
0.65
0.57
0.32
0.93
0.92
0.82
0.47
1.02
1.01
0.90
0.52
0.90
0.90
0.80
0.46
0.73
0.73
0.64
0.37
0.53
0.67
0.76
0.57
0.57
0.73
0.84
0.65
0.53
0.67
0.76
0.57
0.34
0.44
0.50
0.37
0.49
0.63
0.72
0.54
0.54
0.69
0.79
0.60
0.48
0.61
0.70
0.53
0.39
0.50
0.57
0.43
V2
V3
V4
V5
V6
V7
V8
maximal ­WPI rather than crop yield, under deficit irrigation conditions (Chen et al. 2009). The results of the study
indicated that the calibrated model was able to effectively
simulate the crop growth under water deficit conditions. It
means that a calibrated crop growth model combined with
an opmization algorithm achieves maximum W
­ PI. Also, it
is suggested that management of rice irrigation is done on
saline soil and under irrigated saline regimes using SWAP.
References
Ahmadzadeh H, Morid S, Delavar M, Srinivasan R (2015) Using
the SWAT model to assess the impacts of changing irrigation
from surface to pressurized systems on water productivity and
13
water saving in the Zarrineh Rud catchment. Agric Water Manag
10.1016/jagwat201510026
Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration-guidelines for computing crop water requirements. FAO
Irrigation and Drainage Paper 56, Rome, Italy, p 299
Amiri E, Rezaei M, Eyshi Rezaei E, Bannayan M (2014) valuation of
Ceres-Rice, Aquacrop and Oryza 2000 models in simulation of
rice yield response to different irrigation and nitrogen management strategies. J Plant Nutr 37:1749–1769
Bessembinder JJE, Leffelaar PA, Dhindwal AS, Ponsioen TC (2005)
Which crop and which drop, and the scope for improvement of
water productivity. Agric Water Manag 73:113–130
Braden H (1985) Ein energiehaushalts- und verdunstungsmodell for
wasser und stoffhaushaltsuntersuchungen landwirtschaftlich
genutzer Einzugsgebiete. Ittelungen Deutsche Bodenkundliche
Geselschaft 42:294–299
Chen C, Wang E, Yu Q (2009) Modeling the effects of climate variability and water management on crop water productivity and
Paddy Water Environ
water balance in the North China Plain. AgricWater Manage.
doi:10.1016/jagwat200811012
Confalonieri R, Rosenmund A, Baruth B (2009) An improved model
to simulate rice yield. Agron Sustain Dev 29:463–474
Droogers P, Bastiaanssen WGM (2002) Irrigation performance using
hydrological and remote sensing modeling. J Irrig Drain Eng
128:11–18
Droogers P, Akbari M, Torabi M, Pazira E (2000) Exploring field scale
salinity using simulation modeling, example for Rodasht area,
Esfahan Province, Iran. IAERIIWMI Research Reports 2
English M (1990) Deficit irrigation. I: analytical framework. Irrig
Drain E-ASCE 116:399–412
Feddes RA, Kowalik PJ, Zaradny H (1978) Simulation of field water
use and crop yield. Simulation monographs. Pudoc, Wageningen,
p 188
Govindarajan S, Ambujam NK, Karunakaran K (2008) Estimation
of paddy water productivity (WP) using hydrological model: an
experimental study. Paddy Water Environ, 6:327–339
Ines AV, Honda K, Das Gupta A, Droogers P, Clemente RS (2006)
Combining remote sensing-simulation modeling and genetic
algorithm optimization to explore water management options in
irrigated agriculture. Agric Water Manag 83:221–232
Jiang Y, Xua X, Huang Q, Huo Z, Huang G (2015) Assessment of irrigation performance and water productivity in irrigated areas of the
middle Heihe River basin using a distributed agro-hydrological
model. Agric Water Manag 147:67–81
Jiang J, Feng S, Ma J, Huo Z, Zhang C (2016) Irrigation management
for spring maize grown on saline soil based on SWAP model.
Field Crop Res 196:85–97
Keller A, Seckler D (2005) Limits to the productivity of water in crop
production. California Water Plan Update 4:178–197
Krause P, Boyle DP, Ase FB (2005) Comparison of different efficiency
criteria for hydrological model assessment. Adv Geosci 5:89–97
Kroes JG, Van Dam JC, Groenendijk P, Hendriks RFA, Jacobs CMJ
(2008) SWAP version 3.2. Theory description and user manual.
Wageningen, Alterra, Alterra Report 1649
Lane L, Ferreira V (1980) Sensitivity analysis. In: Knisel WG (ed.)
CREAMS: a sield-scale model for chemicals, runoff, and erosion
from agricultural management systems. Cons Res Rpt No 26 US
Dept Agr, Washington, DC, pp 13–35
Liu J, Willium JR, Zehnder AJB, Yang H (2007) GEPIC- modelling
wheat yield and crop productivity with high resolution on a global
scale. Agric Syst 94:478–493
Loeve R, Barker R, Dawe D, Lin H, Bin D (2003) Growing more rice
with less water: an overview of research in liuyuankou irrigation
system, Henan Province, China. Int Water Manag Inst (IWMI)
73:113–130
Molden D, Sakthivadivel R (1999) Water accounting to assesses and
productivity of water. Water Resour Dev 15:55–72
Nash JE, Sutcliffe JV (1970) River flow forecasting through conceptual
models. A discussion of principles. J Hydrol 10:282–290
Noory H, Van der Zee SEATM, Liaghat AM, Parsinejad M, van Dam
JC (2011) Distributed agro-hydrological modeling with SWAP
to improve water and salt management of the Voshmgir irrigation and drainage network in Northern Iran. Agric Water Manag
98:1062–1070
Singh R, van Dam JC, Feddes RA (2006) Water productivity analysis of irrigated crops in Sirsa district. India Agric Water Manag
82:253–278
Van Dam JC et al (1997) Theory of SWAP version 2.0. Department
of water resour., Report 71. Wageningen Agricultural University,
Wageningen, p 167
Van Dam JC, Groenendijk P, Hendriks RFA, Kroes JG (2008)
Advances of modeling water flow in variably saturated soils with
SWAP. Vadose Zone J 7:640–653
Van der Knijff JM, Younis J, De Roo APJ (2010) a GIS-based distributed model for river basin scale water balance and flood simulation. Int J Geogr Inf Sci 24:189–212
Van Genuchten MT (1980) A closed form equation for predicting the
hydraulic conductivity of unsaturated soils. Soil Sci Soc Am J
44:892–898
Vazifedoust M, Van Dama JC, Feddes RA, Feizi M (2008) Increasing
water productivity of irrigated crops under limited water supply
at field scale. Agric Water Manag 95:89–102
Vermaa AK, Gupta SK, Isaac RK (2012) Use of saline water for irrigation in monsoon climate and deep water table regions: simulation
modeling with SWAP. Agric Water Manag 115:186–193
Vilayvong S, Banterng P, Patanothai A, Pannangpetch K (2015) CSMCERES-Rice model to determine management strategies for lowland rice production. Sci Agric 72:229–236
Walkley A, Black IA (1934) An examination of Degtjareff method for
determining soil organic matter and a proposed modification of
the chromic acid titration method. Soil Sci 37:29–37
Wösten JHM, Lilly A, Nemes A, Le BC (1998) Using existing soil data
to derive hydraulic parameters for simulation models in environmental studies and in land use planning. Report 156 DLO, Winand
Staring Centre The Netherlands
Zare N, Khaledian M, Pirmoradian N, Rezaei M (2014) Simulation
of rice yield under different irrigation and nitrogen application
managements by CropSyst model. Acta Agric Slov 103:181–190
Zwart SJ, Bastiaanssen WGM (2004) Review of measured crop water
productivity values for irrigated wheat, rice, cotton and maize.
Agric Water Manag 69:115–133
13
Документ
Категория
Без категории
Просмотров
1
Размер файла
1 687 Кб
Теги
0622, 017, s10333
1/--страниц
Пожаловаться на содержимое документа