ORIGINAL PAPER

Application of the CSM-CERES-Rice model for evaluationof plant density and irrigation management of transplanted ricefor an irrigated semiarid environment

Shakeel Ahmad • Ashfaq Ahmad • Hakoomat Ali • Abid Hussain •

Axel Garcia y Garcia • Muhammad Azam Khan • Muhammad Zia-Ul-Haq •

Mirza Hasanuzzaman • Gerrit Hoogenboom

Received: 23 August 2011 / Accepted: 1 February 2012 / Published online: 23 February 2012

� Springer-Verlag 2012

Abstract The objectives of this study were to evaluate

the performance of the Cropping System Model (CSM)-

CERES (Crop-Environment Resource Synthesis)-Rice for

simulating growth and yield of rice under irrigated condi-

tions for a semiarid environment in Pakistan and to deter-

mine the impact of plant density and irrigation regime on

grain yield and economic returns. The crop simulation

model was evaluated with experimental data collected

in 2000 and 2001 in Faisalabad, Punjab, Pakistan. The

experiment utilized a randomized complete block design

with three replications and included three plant densities

(one seedling hill-1, PD1; two seedlings hill-1, PD2; and

three seedlings hill-1, PD3) and five irrigation regimes

(625 mm, I1; 775 mm, I2; 925 mm, I3; 1075 mm, I4; and

1225 mm, I5). To determine the most appropriate combi-

nation of plant densities and irrigation regimes, four plant

densities from one seedling hill-1 to four seedlings hill-1

and 17 irrigation regimes ranging from 0 to 1600 mm, for a

total of 68 different scenarios, were simulated for 35 years

of historical daily weather data. The evaluation of CSM-

CERES-Rice showed that the model was able to accurately

simulate growth and yield of rice for irrigated semiarid

conditions, with an average error of 11% between simu-

lated and observed grain yield. The results of the bio-

physical analysis showed that the combination of the two

seedlings hill-1 plant density and the 1,300 mm irrigation

regime produced the highest yield compared to all other

scenarios. Furthermore, the economic analysis through the

Mean-Gini Dominance (MGD) also showed the superiority

of this treatment compared to the other treatment combi-

nations. The mean monetary return ranged from -47 to

1,265 $ ha-1 among all 68 scenarios. However, to be able

to furnish the demand of rice grain for local consumption

and to increase export, there is a need to expand this

technology among the rice growers of other rice producing

areas in Pakistan through extension workers.

Introduction

Irrigated agriculture is the largest freshwater-consumption

sector but it faces competing demands from other sectors,

such as industry and domestic use (Zwart and Bastiaanssen

2004). Agriculture uses 70% of all freshwater worldwide

Communicated by S. Azam-Ali.

S. Ahmad � H. Ali

Department of Agronomy, Bahauddin Zakariya University,

Multan 60800, Pakistan

A. Ahmad � A. Hussain

Agro-climatology Lab, University of Agriculture,

Faisalabad 38040, Pakistan

A. Garcia y Garcia

Department of Plant Sciences, University of Wyoming,

Powell, WY 82435-9135, USA

M. A. Khan

Executive District Officer Agriculture, Chiniot,

Punjab 35400, Pakistan

M. Zia-Ul-Haq

Department of Pharmacognosy, University of Karachi,

Karachi 75270, Pakistan

M. Hasanuzzaman (&)

Department of Agronomy, Faculty of Agriculture,

Sher-e-Bangla Agricultural University, Dhaka 1207, Bangladesh

e-mail: mhzsauag@yahoo.com

G. Hoogenboom

AgWeatherNet Program, Washington State University,

Prosser, WA 99350-8694, USA

123

Irrig Sci (2013) 31:491–506

DOI 10.1007/s00271-012-0324-6

(Kunimitsu 2009), while rice (Oryza sativa L.) production

consumes about 30% (Peng et al. 2006). In Asia, rice

consumes more than 45% of total freshwater (Barker et al.

1999). To meet the rising demand for food security due to

the increase in population and economic development, rice

yield should increase by more than 1.2% per year world-

wide (Normile 2008). However, the scarcity of freshwater

resources has threatened the production of flooded rice

(IWMI 2000). It is expected that by 2025, 15 out of 75

million ha of Asia’s flooded rice will experience water

shortage (Tuong and Bouman 2003). Therefore, there is

increasing pressure to reduce freshwater use in irrigated

agriculture through alternative conservation practices. A

major challenge for agronomists across the world is the

development of technologies for increased and sustained

rice productivity with the currently available water

resources (Zeigher and Puckridge 1995). Anaerobic

(flooded ? puddled) and aerobic (non-flooded and non-

puddled) are two common rice production systems in the

world, including in Asia. There is substantial water savings

in the production of aerobic rice compared to anaerobic

rice due to lower losses through seepage, percolation, and

evaporation (Bouman et al. 2002; Tuong et al. 2005; Peng

et al. 2006). Therefore, many water-saving management

technologies have been developed for flooded rice to

reduce water losses and to increase water productivity.

These water-saving practices include saturated soil culture

(Borell et al. 1997; Bouman and Tuong 2001), alternate

wetting and drying (Li 2001; Bouman and Tuong 2001;

Tabbal et al. 2002; Soundharajan and Sudheer 2009), and

ground cover systems (Lin et al. 2002; Feng et al. 2007).

The amount of irrigation water saved and the yield that can

be obtained using water-saving technologies vary greatly

among climate, soil type, hydrological conditions, and rice

production systems (Cabangon et al. 2004; Yang et al.

2005; Feng et al. 2007; Soundharajan and Sudheer 2009).

Irrigation water conservation technologies are applica-

ble and effective only when there is an assured supply of

irrigation water. In Pakistan, the water availability for

canals is shrinking due to decreasing storage capacity of

dams and decreasing rainfall. In addition, farmers lack

access to water for irrigation on their individual farms as

the management of surface irrigation is based on weekly

turns. In addition, the present system lacks consideration of

rainfall or droughts and continues to deliver the same

amount of water as scheduled (Choudhary 2008). Even

during a farmer’s turn, water is only available depending

upon the area of their land holding rather than the actual

amount of water required for a particular crop or area to be

irrigated (GOP 2008). The allocated water times usually

vary from 30 to 55 min ha-1 depending upon distance

from the distributary canal, minor canals, or water courses

(GOP 2009). The alternative to canal water is groundwater,

which is not sustainable for resource-poor farmers due to

the increasing prices of electric power and diesel that are

required to operate the pumps.

In Pakistan, rainfall is not only scarce but also unevenly

distributed (GOP 2007, 2009). Most of the rainfall occurs

during the monsoon season in July and August with heavy

floods after drought conditions have occurred. In Pakistan

and other countries in the region, rainfall is barely suffi-

cient to support a one-season rice crop (GOP 2009).

Therefore, production relies heavily on surface and ground

water (Bouman and Tuong 2001; Geng et al. 2001; Feng

et al. 2007). The only option, therefore, is to grow irrigated

rice in Pakistan. In this study, irrigated rice refers to the

non-flooded lowland rice with supplemental irrigation. The

primary potential threat to rice productivity in Pakistan is

water shortage. However, optimum plant population should

be ensured to achieve higher yields. A plant density either

below or above the optimum affects yield due to a smaller

number of plants per unit area or inter-plant competition at

higher plant densities (Ahmad et al. 2011). However,

judicious use of water with respect to the time and rate of

application according to the need of the crop is very

important. Irrigated rice in Pakistan is an imperative target

for water conservation when compared to other crops

having less water requirements.

Crop simulation models linked to decision support sys-

tems have been used successfully for a wide range of

applications in many countries around the world to assess

alternative crop management options (Tsuji et al. 1994,

1998; Jones et al. 2003; Hoogenboom et al. 2010). The

Decision Support System for Agro-technology Transfer

(DSSAT) is a widely used decision support system that

includes the Cropping System Model (CSM) (Jones et al.

2003; Hoogenboom et al. 2010). CSM is a process-ori-

ented, dynamic crop simulation model that stimulates crop

growth, development, and yield for 27 food and other

crops, including rice (Jintrawet 1995; Ritchie et al. 1998;

Cheyglinted et al. 2001; St’astna et al. 2002; Gungula et al.

2003; Kumar and Sharma 2004; Sarkar and Kar 2006;

Ahmad et al. 2011). The CSM-CERES-Rice model simu-

lates growth and development from either sowing or

transplanting to harvest maturity and is based on the

physiological processes that describe the response of rice to

local soil and weather conditions. Potential growth is

dependent upon photosynthetically active radiation, light

interception, and light conversion efficiency, where as

biomass production on any day is constrained by crop

management, suboptimal temperature, soil water deficit,

nitrogen deficiency, and their respective interactions. The

input data required to run the DSSAT models include daily

weather data, i.e., maximum and minimum temperature,

rainfall, and solar radiation; soil characterization data, e.g.,

physical, chemical, and morphological properties for each

492 Irrig Sci (2013) 31:491–506

123

soil horizon; genetic information through cultivar coeffi-

cients; and crop management information such as trans-

planting date, age of nursery-transplanted seedlings, row

and plant spacing, rates, dates and amounts of fertilizer and

irrigation application. The model calculates the soil water

balance of the rice crop on a daily basis as a function of

precipitation, irrigation, soil surface runoff and drainage

from the bottom of the profile, and transpiration and soil

evaporation (Ritchie 1998). The model also calculates the

soil nitrogen balance as a function of a range of soil

nitrogen transformation processes as well as nitrogen

uptake by the plant (Godwin and Singh 1998; Ahmad et al.

2011).

Most of the studies during the past decade have used

crop simulation models with single or multiple factors but

without considering the economic impact for growers

(Cheyglinted et al. 2001; Kropff et al. 2001; Timsina and

Connor 2001; St’astna et al. 2002; Timsina and Humphreys

2006; Arora 2006; Soler et al. 2007a, b; Banterng et al.

2010). Limited scientific information is available regarding

the application of the CSM-CERES-Rice for the interaction

of multiple crop management factors such as plant density

and irrigation regime for irrigated conditions in a semiarid

environment. The overall goal of this study, therefore, was

to identify the optimum management practices for rice

using a combination of field experiments and crop simu-

lation models. The specific objectives of this study

included (1) to evaluate of the performance of the CSM-

CERES-Rice model for plant density and irrigation man-

agement under semiarid conditions and (2) to determine the

best management options for increasing rice productivity

for local conditions in Pakistan.

Materials and methods

Experimental location and design

The CSM-CERES-Rice model of DSSAT Version 4.5

(Jones et al. 2003; Hoogenboom et al. 2010), including the

improved soil water model (Ritchie et al. 2009), was

evaluated with experimental data that were collected in

Faisalabad, Pakistan. The observed data were obtained

from two experiments conducted at the University of

Agriculture, Faisalabad, Pakistan in 2000 and 2001. A

randomized complete block design (RCBD) was employed

with three replications. An aromatic variety, i.e., Basmati-

385, was grown during both years of the experiment. Daily

weather records for the experimental site (just at 200

meters distance) were obtained from the Department of

Crop Physiology, University of Agriculture Faisalabad

(UAF), (36.25� lat N, 73.09� long E, and 184.4 m altitude

from sea level) Pakistan. The soil was a Lyallpur clay loam

(aridisol-fine-silty, mixed, hyperthermic Ustalfic, Hap-

larged [USDA classification] and Haplic Yermosols [FAO

classification]). The soil was silt–clay loam in nature with

soil texture of sand 51.31%, silt 21.04%, and clay 27.65%

at 0–15 cm, while respective values for 16–30 cm were

52.05, 20.83 and 27.12, respectively. During both years,

the previous crop was wheat. Prior to sowing, soil analysis

at a depth of 0–15 cm showed a pH of 7.8, total soluble

salts at 0.22%, organic matter at 0.76%, total nitrogen at

0.046%, available phosphorus at 6.19 ppm, and available

potassium at 193 ppm. The values for 16–30 cm were 7.9,

0.23, 0.73, 0.048, 6.11, and 192, respectively.

Experimental procedures

Thirty-day-old seedlings were transplanted manually in a

puddled field in standing water at a plant distance of

22.5 cm and a row distance of 22.5 cm for both years.

Nitrogen was applied in two equal splits through urea.

The first dose (75 kg N ha-1) was applied during pud-

dling on July 4, 2000 and July 3, 2001 prior to the final

cultivation. The second dose (75 N kg ha-1) was applied

at 22 days after transplanting (DAT) on July 26, 2000 and

July 25, 2001. In addition, phosphorus and potassium

(P2O5 and K2O; 65 kg ha-1 each) in the form of single

super phosphate (SSP), and potassium sulfate and zinc at

a rate of 25 kg ha-1 were also applied during puddling

prior to final cultivation. All other agronomic practices

such as weeding and plant protection measures were

standard and uniform for all the treatments. Additional

details on crop management, plant sampling and obser-

vations can be found in Ahmad et al. (2005a, b, 2008,

2009a, b, 2011).

For irrigation, water was applied at fixed weekly inter-

vals by adding the approximately 77 mm of water. This

fixed amount was applied regardless of rainfall, because

during the rice growing season, the temperature is very

high in this region and rainfall is highly unpredictable in

terms of intensity and the amount of rain. In addition,

farmers do not have their own tube well and the farmers’

turn from the canal source is weekly and is independent of

local rainfall as was explained in the introduction. To

emulate these local restrictions, we applied water on a

weekly basis. The experimental plots were irrigated

according to the different irrigation treatments. Water in

the amount specified by each treatment was applied using a

cutthroat flume (90 cm 9 20 cm). The duration for each

irrigation application was calculated using the following

Eq. 1

T ¼ A� d=Q; ð1Þ

where T is the time required for application of irrigation

water to an area (A) to be irrigated at depth (d); and Q is

Irrig Sci (2013) 31:491–506 493

123

discharge or flow rate. The experiment included the

following treatment combinations: three plant densities,

viz., one seedling hill-1 (PD1), two seedlings hill-1

(PD2), and three seedlings hill-1 (PD3) and five irrigation

regimes, viz., 625 mm (eight irrigations, I1), 775 mm

(ten irrigations, I2), 925 mm (twelve irrigations, I3),

1075 mm (fourteen irrigations, I4), and 1225 mm (six-

teen irrigations; I5). For all irrigation treatments, water

Table 1 Genetic coefficients of the CSM-CERES-Rice model for the fine rice cultivar Basmati-385

Cultivar traits Genetic coefficients Unit Value

Vegetative growth

1. Time from seed emergence to the end of juvenile phase P1 Photothermal day 126.0

2. Extent to which development is delayed for each hour increase

in photoperiod above the longest photoperiod

P2O h 250.0

3. Extent to which phasic development from vegetative to panicle

initiation is delayed for each hour increase in photoperiod above P2O, i.e. 12.5 h

P2R Photothermal day 500.0

Reproductive growth

4. Time starting from grain filling to physical maturity P5 Photothermal day 10.9

5. Maximum spikelet number coefficient G1 – 52.0

6. Maximum possible single grain size G2 g 0.0235

7. Scalar vegetative growth coefficient for tillering relative to IR64 G3 – 1.03

8. Temperature tolerance scalar coefficient G4 – 0.90

Source: Ahmad et al. (2011)

a

Tem

pera

ture

(° C

)

0

10

20

30

40

50b

d

140 160 180 200 220 240 260

Sol

ar r

adia

tion

(M

J m

-2 d

ay-1

)

0

5

10

15

20

25

30

Maximum temperatureMinimum temperature

c

Day of year

140 160 180 200 220 240 260

Pre

cipi

tati

on (

mm

)

0

20

40

60

PrecipitationSolar radiation

2000 2001

Fig. 1 Daily weather

conditions for Faisalabad,

Pakistan, during a typical rice

growing season; maximum and

minimum air temperature for

2000 (a) and 2001 (b) and solar

radiation and precipitation for

2000 (c) and 2001 (d)

494 Irrig Sci (2013) 31:491–506

123

was applied weekly but irrigation was ended at different

dates. For 2002, these irrigation termination dates were

23 August, 6 September, 20 September, 4 October, and

18 October for the irrigation regimes I1, I2, I3, I4, and I5,

respectively. For 2001, these dates were 22 August, 5

September, 19 September, 3 October, and 17 October,

respectively.

Calibration and evaluation of the CSM-CERES-Rice

model

The CSM-CERES-Rice model was calibrated with the data

obtained from the 2000 field experiment for the treatment

that consisted of a plant density of two seedlings hill-1 and

a irrigation regime of 1225 mm as it had the best perfor-

mance. The cultivar coefficients of the variety Basmati-385

are presented in Table 1. The performance of the model,

including its response to irrigation, was evaluated with the

data obtained from the remainder of the treatments of the

2000 and 2001 field experiments. As part of the evaluation

process, the simulated data for LAI, above-ground bio-

mass, and final grain yield were compared with the

observed values from both experiments. The statistical

indices that were used for model evaluation included the

root mean square error (RMSE) (Wallach and Goffinet

1987) and the index of agreement (d-stat) (Willmott 1982;

Willmott et al. 1985). The RMSE was calculated according

to Eq. 2

Root mean square errorðRMSEÞ ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

P

n

i¼1

Pi� Oið Þ2

n

v

u

u

u

t

2

6

6

6

4

3

7

7

7

5

;

ð2Þ

where Pi and Oi refer to the predicted and observed values

for the studied variables, respectively, e.g., leaf area index,

total biomass, and grain yield, and n is the mean of the

observed variables. Normalized RMSE (RMSEn) gives a

measure (%) of the relative difference of simulated versus

observed data. The RMSEn was calculated following Eq. 3

Normalized root mean square error ¼ RMSE� 100

O

� �

ð3ÞThe index of agreement (d) proposed by Willmott et al.

(1985) was estimated using Eq. 4

Index of agreement dð Þ ¼ 1�

P

n

i¼1

Pi� Oið Þ2

P

n

i¼1

ð P0ij j þ O0ij jÞ2

2

6

6

4

3

7

7

5

;

0� d� 1;

ð4Þ

where n is the number of observations, Pi the predicted

observation, Oi is a measured observation and P0i = Pi -

M and O0i = Oi - M (M is the mean of the observed

variable). The percentage deviations (PD) were also

Table 2 Average simulated (S) and observed (O) phenology, root

mean square error and d-value for growth characteristics from CSM-

CERES-Rice model calibration (2000) and evaluation (2001) for the

fine rice cultivar Basmati-385 with three plant densities and five

irrigation levels (n = 15)

Transplanting date Crop characteristics

Phenology S (DAT) O (DAT

July 5, 2000 Anthesis date 72 72

Maturity date 102 103

Growth RMSE d-value

July 5, 2000 LAI (m2 m-2) 1.12 0.95

Top weight (kg ha-1) 365 0.97

Phenology S (DAT) O (DAT

July 4, 2001 Anthesis date 73 72

Maturity date 103 103

Growth RMSE d-value

July 4, 2001 LAI (m2 m-2) 1.29 0.96

Top weight (kg ha-1) 385 0.98

DAT, days after transplanting; LAI, leaf area index

Irrig Sci (2013) 31:491–506 495

123

calculated. The computed values of RMSE and d determine

the degree of agreement between the predicted values with

their respective observed values. A low value for RMSE

and a d-value that approaches 1 are desirable.

Model application

Biophysical analysis

The CSM-CERES-Rice model was used for long-term

simulations of the aromatic rice cultivar Basmati-385. The

CERES-Rice model includes an automatic irrigation

management option that applies irrigation when certain soil

moisture conditions are met. This includes an irrigation

threshold value that sets the available soil moisture for a

certain depth; both can be defined by the user. The simu-

lations were conducted for 35 years using observed daily

weather data starting in 1974 and ending in 2008. The

seasonal analysis program of DSSAT v.4.5 was used to

evaluate and compare different combinations of crop

management practices (Thornton and Hoogenboom 1994;

Thornton et al. 1998; Hoogenboom et al. 2010). The

overall goal was to determine the best management option

for a rice crop grown under irrigated semiarid conditions

similar to the environment of Faisalabad, Pakistan. The

scenarios that were evaluated included combinations of

a

0.0

0.5

1.0

1.5

2.0

2.5b c

02000400060008000100001200014000

d

0.0

0.5

1.0

1.5

2.0

2.5e f

02000400060008000100001200014000

g

LAI (

m2

m-2

)

0.0

0.5

1.0

1.5

2.0

2.5h i

Bio

mas

s (kg

dm

ha-1

)

02000400060008000100001200014000

j

0.0

0.5

1.0

1.5

2.0

2.5k l

02000400060008000100001200014000

m

0 20 40 60 80 1000.0

0.5

1.0

1.5

2.0

2.5

Observed LAI Simulated LAIObserved biomass Simulated biomass

n

Days after transplanting0 20 40 60 80 100

o

0 20 40 60 80 10002000400060008000100001200014000

Fig. 2 Simulated (continuousand dotted lines) and observed

(triangular and round symbols)

leaf area index and biomass of

rice Basmati-385 at variable

plant densities and irrigation

levels; PD1 ? I1 (a), PD2 ? I1

(b), PD3 ? I1 (c), PD1 ? I2 (d),

PD2 ? I2 (e), PD3 ? I2 (f),PD1 ? I3 (g), PD2 ? I3 (h),

PD3 ? I3 (i), PD1 ? I4 (j),PD2 ? I4 (k), PD3 ? I4 (l),PD1 ? I5 (m), PD2 ? I5 (n) and

PD3 ? I5 (o) under irrigated

semiarid conditions at

Faisalabad, Pakistan, for 2000

(see the ‘‘Materials and

methods’’ section for the

definition of the individual

treatment combinations)

496 Irrig Sci (2013) 31:491–506

123

four plant densities and 17 irrigation regimes for a total of

68 scenarios (treatment combinations). The simulation

results were analyzed using the strategy analysis program

of DSSAT (Thornton et al. 1995, 1998) and comparing

percentile distributions for grain yield and water use.

Economic analysis

The results of each strategy, i.e., each combination of plant

density and irrigation regimes, were also evaluated using the

Mean-Gini Dominance Analysis (Buccola and Subaei 1984;

Fawcett and Thornton 1989), an evaluation procedure of the

seasonal analysis program that calculates the monetary return

for each treatment combination and identifies the most dom-

inant treatment based on the highest economic return (Gini

coefficient) and comparing percentile distributions for mon-

etary return. The gross margin ($ ha-1) for each of treatment

was determined using the following equation

GM ¼ Y � P� I � C � V ð5Þ

a

0.0

0.5

1.0

1.5

2.0

2.5b c

0

2000

4000

6000

8000

10000

12000

14000

d

0.0

0.5

1.0

1.5

2.0

2.5e f

0

2000

4000

6000

8000

10000

12000

14000

g

LA

I (m

2 m-2

)

0.0

0.5

1.0

1.5

2.0

2.5h i

Bio

mas

s (k

g dm

ha-

1 )

0

2000

4000

6000

8000

10000

12000

14000

j

0.0

0.5

1.0

1.5

2.0

2.5k l

0

2000

4000

6000

8000

10000

12000

14000

m

0 20 40 60 80 1000.0

0.5

1.0

1.5

2.0

2.5

Observed LAI Simulated LAIObserved biomass Simulated biomass

n

Days after transplanting0 20 40 60 80 100

o

0 20 40 60 80 1000

2000

4000

6000

8000

10000

12000

14000

Fig. 3 Simulated (continuousand dotted lines) and observed

(triangular and round symbols)

leaf area index and biomass of

rice Basmati-385 at variable

plant densities and irrigation

levels, i.e., PD1 ? I1 (a),

PD2 ? I1 (b), PD3 ? I1 (c),

PD1 ? I2 (d), PD2 ? I2 (e),

PD3 ? I2 (f), PD1 ? I3 (g),

PD2 ? I3 (h), PD3 ? I3 (i),PD1 ? I4 (j), PD2 ? I4 (k),

PD3 ? I4 (l), PD1 ? I5 (m),

PD2 ? I5 (n) and PD3 ? I5

(o) under irrigated semiarid

conditions at Faisalabad,

Pakistan, for 2001 (see the

‘‘Materials and methods’’

section for the definition of the

individual treatment

combinations)

Irrig Sci (2013) 31:491–506 497

123

Table 3 A comparison of simulated (S) and observed (O) grain yield for the fine rice cultivar Basmati-385 as affected by plant density and

irrigation levels at Faisalabad, Pakistan

Irrigation levels Grain yield

(kg ha-1)

PD (%)a GY d-valueb RMSEn (%)c RMSE

(kg ha-1)dMSEe MSEsf MSEug LAI

d-valuehBiomass

d-valuei

Sim. Obs.

2000

One seedling hill-1

625 mm (I1) 3445 3062 10.9 0.92 0.96

775 mm (I2) 4152 3685 11.2 0.95 0.96

925 mm (I3) 4428 4081 7.8 0.92 0.96

1075 mm (I4) 4629 4307 7.0 0.94 0.96

1225 mm (I5) 4815 4379 9.1 0.95 0.95

Statistics 0.94 2.1 395 155793 106765 49028 1.12 140

Two seedlings hill-1

625 mm (I1) 3627 3664 1.2 0.94 0.94

775 mm (I2) 4338 3915 9.8 0.96 0.97

925 mm (I3) 4677 4305 8.0 0.95 0.96

1075 mm (I4) 4863 4647 4.4 0.92 0.97

1225 mm (I5) 5265 4712 10.5 0.96 0.97

Statistics 0.97 1.8 366 134229 91987 42242 1.14 137

Three seedlings hill-1

625 mm (I1) 3968 3503 4.1 0.92 0.95

775 mm (I2) 4053 3798 6.2 0.94 0.97

925 mm (I3) 4209 4074 3.3 0.93 0.96

1075 mm (I4) 4882 4519 7.4 0.94 0.96

1225 mm (I5) 5171 4604 11.1 0.97 0.98

Statistics 0.96 1.4 333 111166 76182 34984 1.08 138

2001

One seedling hill-1

625 mm (I1) 3945 3550 10.0 0.94 0.97

775 mm (I2) 4352 3880 10.8 0.96 0.98

925 mm (I3) 4729 4300 9.1 0.94 0.97

1075 mm (I4) 4833 4540 6.1 0.95 0.97

1225 mm (I5) 5015 4630 7.7 0.96 0.96

Statistics 0.96 2.1 399 159385 109227 50158 1.33 174

Two seedlings hill-1

625 mm (I1) 3792 3840 1.3 0.92 0.96

775 mm (I2) 4565 4110 9.9 0.95 0.96

925 mm (I3) 4877 4540 6.9 0.96 0.98

1075 mm (I4) 5063 4900 3.2 0.96 0.97

1225 mm (I5) 5474 4970 9.2 0.97 0.98

Statistics 0.99 1.7 347 120697 82714 37983 1.29 165

Three seedlings hill-1

625 mm (I1) 4083 3690 9.6 0.94 0.96

775 mm (I2) 4372 4000 8.5 0.95 0.98

925 mm (I3) 4512 4300 4.7 0.96 0.97

1075 mm (I4) 4995 4750 4.9 0.95 0.98

1225 mm (I5) 5386 4860 9.8 0.96 0.99

Statistics 0.97 1.6 367 134896 99444 45452 1.26 171

a Percentage prediction deviation; b,h,i d-value; c Normalized root mean square error; d Root mean square error; e Mean square error;f Systematic MSE; g Unsystematic MSE

498 Irrig Sci (2013) 31:491–506

123

, where Y is the simulated rice grain yield (kg ha-1), P is

the price of rice ($ t-1) (average for the past 3 years,

2007–2009), I is the irrigation application rate ($ applica-

tion-1) as per treatment, C is the cost of irrigation

($ mm-1), and V is the base production cost ($ ha-1) for

all treatments during 2009. The price (297 $ t-1), cost of

irrigation (0.47 $ mm-1), and base production cost

(275 $ ha-1) for rice was used in this study. The base

production cost for rice was obtained from the Department

of Agriculture Extension, Punjab, Pakistan (unpublished

data). The rice prices were obtained from the Economic

Survey of Pakistan, Finance Division (available online)

(GOP 2009).

Weather conditions for Faisalabad

The climate of the experimental site is semiarid with an

average annual maximum and minimum temperature of

31.2 and 17.2�C, respectively, and annual rainfall that

ranges from 400 to 800 mm. However, during the rice

growing season which lasts from May through October,

the average seasonal maximum and minimum air tem-

peratures are 37.1 and 24.3�C, respectively. Overall, 75%

of the annual precipitation occurs during the rice growing

season (GOP 2009). The monthly average solar radiation

during the rice growing season ranges from 20 to

24.5 MJ m-2 d-1. A summary of the weather data for the

two rice cropping seasons of 2000 and 2001 are presented

in Fig. 1.

Results and discussion

Model calibration and evaluation

The CSM-CERES-Rice model was calibrated with the data

obtained from the 2000 field experiment for the treatment

that consisted of a plant density of two seedlings hill-1 and

a nitrogen application of 200 kg N ha-1 as it had the best

performance. The cultivar coefficients for the rice cultivar

Basmati-385 were estimated through trial and error and a

comparison between model-simulated and experimental

data. The final values for the eight cultivar coefficients that

determine vegetative (P1, P2O, and P2R) and reproductive

(P5, G1, G2, G3, and G4) growth and development are

presented in Table 1. After the calibration of the cultivar

coefficients, the accuracy of the model was checked with

observed data for the remaining 14 treatments for 2000. As

part of the calibration and evaluation process, the simulated

data for anthesis date, maturity date, grain yield, and total

biomass were compared with the observed values. A close

agreement was obtained between observed and simulated

values for rice phenology. For 2000, the model predicted

the dates from transplanting to anthesis and transplanting to

maturity with a zero difference between observed and

simulated dates for cultivar Basmati-385 (Table 2). The

simulated and observed values were in good agreement for

LAI and total above-ground biomass at different pheno-

logical stages for the different combinations of plant den-

sity and nitrogen rates (Fig. 2). The average index of

agreement (d-value) across all treatments for LAI was 0.95,

while the RMSE was 1.12 m2 m-2 (Table 2). The d-value

for above-ground biomass was 0.97, while the RMSE was

365 kg ha-1 (Table 2). The experimental data collected in

2001 were used for independent model evaluation. For

2001, the model predicted flowering 1 day late, while the

simulated maturity date was predicted on the same date as

the observed maturity. The d-value for LAI was 0.96 and

for top weight was 0.98, slightly better than the statistics

for the calibration, while the RMSE values were 1.29

1.12 m2 m-2 for LAI and 385 kg ha-1 for top weight,

slightly higher than the statistics for the calibration. The

lower values for RMSE and higher d-values close to 1

reflected that the model predicted LAI and above-ground

biomass quite well. The regression line between simulated

and observed grain yield had a value of 0.97 for R2 for the

15 treatment combinations of plant density and irrigation

rates (Fig. 4), showing the ability of model to simulate rice

growth and development in response to different irrigation

levels for a semiarid environment.

The simulated and observed values for the individual

treatments were in good agreement for LAI and total

above-ground biomass at different growth stages for the

different combinations of plant density and irrigation levels

Sim

ulat

ed r

ice

grai

n yi

eld

(kg

ha-1

)

3000

3500

4000

4500

5000

5500

6000

2000 2001

Observed rice grain yield (kg ha-1)

500035003000 000600540004 5500

Fig. 4 Observed and simulated rice grain yields for the fine rice

cultivar-385 at Faisalabad, Pakistan, during 2000 and 2001

Irrig Sci (2013) 31:491–506 499

123

for 2000 and 2001 (Figs. 2, 3). The d-value for LAI ranged

from 0.92 to 0.97 for both years, while RMSE ranged from

1.08 to 1.14 m2 m-2 for 2000 and from 1.26 to

1.33 m2 m-2 for 2001 (Table 3). The d-value for above-

ground biomass ranged from 0.94 to 0.98 for 2000 and

from 0.96 to 0.99 for 2001, while the RMSE ranged from

a-One seedling hill-1

3000

3500

4000

4500

5000

5500

c-Two seedlings hill-1

Sim

ulat

ed a

nd o

bser

ved

rice

gra

in y

ield

(kg

ha-1

)

3000

3500

4000

4500

5000

5500

e-Three seedlings hill-1

Irrigation regimes (mm)

3000

3500

4000

4500

5000

5500

b- One seedling hill-1

d-Two seedlings hill-1

f-Three seedlings hill-1

800 1000 1200 1400 1600

Observed 2000 and 2001Simulated 2000 and 2001

Observed 2000 and 2001Simulated 2000 and 2001

Observed 2000 and 2001Simulated 2000 and 2001

800 1000 1200 1400 1600

Fig. 5 Response curve of both

simulated and observed yield as

a function of total irrigation

applied for the fine rice cultivar

Basmati-385 at Faisalabad

during 2000 (a, c, e) and 2001

(b, d, f)

Table 4 Cumulative irrigation and cumulative irrigation and rainfall (I ? R) for each irrigation regime in relation to days after transplanting at

flowering and maturity for the fine rice cultivar Basmati-385

Irrigation regimes Irrigation water (mm) Cumulative irrigation ? rainfall (mm)

Growth stage

Flowering (72 DAS) Maturity (103 DAS)

2000 2001 2000 2001

I1 = Eight irrigations 625 729 780 742 780

I2 = Ten irrigations 775 879 930 892 930

I3 = Twelve irrigations 925 1,029 1,080 1,042 1,080

I4 = Fourteen irrigations 1,075 1,179 1,230 1,192 1,230

I5 = Sixteen irrigationsa 1,225 1,329 1,380 1,342 1,380

a Current practice

500 Irrig Sci (2013) 31:491–506

123

137 to 140 kg ha-1 for 2000 and from 165 to 174 kg ha-1

for 2001 (Table 3). The lower values of RMSE and higher

d-values close to 1 reflected that the model predicted both

LAI and above-ground biomass fairly accurately. At final

harvest, the simulated values were also in good agreement

with the observed values, and the differences for above-

ground biomass ranged from 1.56 to 15% for 2000 and

2001. The RMSE for grain yield at final harvest ranged

from 333 to 395 kg ha-1 for 2000 and from 347 to

399 kg ha-1 for 2001 (Table 3). The normalized RMSE

varied from 1.4 to 2.1% for 2000 and from 1.6 to 2.1% for

2001, while the d-value ranged from 0.94 to 0.97 for 2000

and 0.96 to 0.99 for 2001 (Table 3). The simulated yield

values were within 11% of observed values. The slight

disparities between observed and simulated values may

have been due to factors other than those considered by the

model, including weeds, diseases, and pests. However,

there differences were not significant and the model was

able to successfully simulate grain yield during both

growing seasons. The regression line between simulated

and observed grain yield had high values for r2 for 2000,

e.g., 0.97 and for 2001, e.g., 0.95, for the 15 treatment

combinations of plant density and irrigation regimes

(Fig. 4). This demonstrated that the model was able to

simulate rice growth, development, and yield under irri-

gated conditions for a semiarid environment quite

0 020 2040 4060 6080 80 100100

a

0200

400

600

800

1000

1200

1400

c

0200

400

600

800

1000

1200

1400

e

Sim

ulat

ed E

T a

nd c

umul

ativ

e ir

riga

tion

+ p

reci

pita

tion

0200

400

600

800

1000

1200

1400

g

0200

400

600

800

1000

12001400

i

0

200

400

600

800

1000

1200

1400

b

d

f

h

j

Days after transplanting

Cummulative irrigation + precipitationSimulated ET

Fig. 6 Cumulative simulated evapotranspiration (ET) and cumulative irrigation plus rain as a function of days after transplanting for irrigation

regimes, i.e., I1 (a), I2 (c), I3 (e), I4 (g) and I5 (i) for 2000 and I1 (b), I2 (d), I3 (f), I4 (h) and I5 (j) for 2001

Irrig Sci (2013) 31:491–506 501

123

accurately. The response curve for yield as a function of

the amount of water applied for irrigation showed a very

similar response for both the measured and simulated

values for both years. It increased linearly with the increase

in the amount of water that was applied (Fig. 5). The water

deficit trend was recorded in the order of the I5 treatments

having the highest water stress followed by I4, I3, I2 and I1,

resulting in a simulated yield decrease from 5171 to

3445 kg ha-1 for 2000 and from 5386 to 3945 kg ha-1 for

2001(Fig. 5). The year 2001 was comparatively wetter than

2000 in terms of total amount of rainfall as total precipi-

tation during years 2000 and 2001 was 117 and 165 mm,

respectively. However, during both years, rainfall was

concentrated only during the vegetative phase of the rice

crop. In 2000, total precipitation was 64, 25 and 28 mm

during the months of July, August and September,

respectively, while in 2001, the amount of rainfall was 108

and 57 mm during the months of July and August, with no

rain in September. The daily global solar radiation ranged

from 7.2 to 28.4 MJ m-2 during vegetative phase and 13.1

to 23.8 MJ m-2 during reproductive phase in 2000, and

respective values for respective phases in 2001 were 7.4 to

30.7 MJ m-2 and 15.3 to 22.1 MJ m-2. However, the

mean daily solar radiation was 21.4 and 21.1 MJ m-2

during years 2000 and 2001, respectively. Cumulative

irrigation and cumulative irrigation ? rainfall (I ? R) for

each irrigation regime in relation to days after transplanting

with key growth stages of flowering and maturity are

presented in Table 4. Cumulative simulated evapotranspi-

ration (ET) and cumulative irrigation ? rainfall against

days after transplanting for years 2000 and 2001 are pre-

sented in Fig. 6, showing the differences between irrigation

treatments and years.

Model application

An analysis to identify the optimum combination of plant

density and irrigation regime for rice production under

irrigated conditions for a semiarid environment in Pakistan

using a systems analysis approach was conducted. The

CSM-CERES-Rice model was used to simulate grain yield

for 68 different scenarios, including combinations of four

plant densities and 17 irrigation regimes for irrigated

conditions for a semiarid environment for 35 years using

One seedling hill-1

0

1000

2000

3000

4000

5000

6000

7000 Two seedlings hill-1

Four seedlings hill-1

Irrigation regimes (mm)

0 200 400 600 800 1000 1200 1400 1600

Three seedlings hill-1

Sim

ulat

ed r

ice

grai

n yi

eld

(kg

ha-1

)

0

1000

2000

3000

4000

5000

6000

7000

0 200 400 600 800 1000 1200 1400 1600

Fig. 7 Simulated grain yield

for the fine rice cultivar

Basmati-385 at variable plant

densities and irrigation regimes.

Whiskers represent 10th and

90th percentile, box plot limitsrepresents the 25th and 75th

percentile, box central linerepresents the median, and

outliers represent the minimum

and maximum values for

simulated grain yield

502 Irrig Sci (2013) 31:491–506

123

historical daily weather data from 1974 to 2008. The plant

densities ranged from one seedling hill-1 to four seedlings

hill-1 and the 17 irrigation regimes ranging from 0 to

1600 mm. The irrigation regimes consisted of different

irrigation applications, ranging from no irrigation (0) to 16

irrigations (16) at weekly intervals. For each irrigation

application, the same amount of water was applied, which

was 100 mm. The main goal was to determine whether

irrigation applications could be eliminated, especially

during the end of the growing season within the current

limitations of the Warabandi system. The simulated rice

yield for the 68 scenarios ranged from 392 to

6821 kg ha-1, showing the range for the lowest of the

lowest and highest of the highest simulated yield. The

simulated rice grain yield for all 68 scenarios is depicted in

Fig. 7, showing the minimum, 10th, 25th percentile,

median, 75th, 90th percentile, and maximum values. The

highest median grain yield was obtained for a plant density

of two seedlings hill-1 and an irrigation regime of

1300 mm with a variation (standard deviation; S.D.) of

1233 kg ha-1. The treatment having interaction of plant

density of two seedlings hill-1 and 1225 mm is now gen-

erally recommended by the local extension department for

rice production for the Faisalabad district. For this treat-

ment, the range of the yield around the median for this

treatment was also smaller than for all other scenarios.

The simulated irrigation water productivity for 68 scenar-

ios that were simulated ranged from 0.25 to 1.17 kg m-3,

showing the range for the lowest of the lowest and highest of

the highest simulated water productivity. The simulated water

productivity for rice for all 68 scenarios is depicted in Fig. 8,

showing the minimum, 10th, 25th percentile, median, 75th,

90th percentile, and maximum values. The highest median

water productivity was obtained for a plant density of two

seedlings hill-1 and an irrigation regime of 1300 mm with a

variation (Standard Deviation; S.D.) of 0.12 kg m-3. The

trends of our study are in line with another study that applied

the ORYZA model (Arora 2006); this study showed that the

mean water-input-based and ET-based productivities ranged

from 0.32 to 1.01 kg m-3 for an irrigated semiarid subtropical

environment in India. Our simulated water productivity for

simulated scenarios values was also similar to the measured

values of water productivity for rice that have been reported

for different countries in south Asia, including 0.39 to

0.64 kg m-3 for Bangladesh (Rashid et al. 2009) and 0.20 to

0.91 kg m-3 for Japan (Kato et al. 2009).

a-One seedling hill-1

Wat

er p

rodu

ctiv

ity

(kg

m-3

)0.2

0.4

0.6

0.8

1.0

1.2

1.4b-Two seedlings hill-1

d-Four seedlings hill-1

0 200 400 600 800 1000 1200 1400 1600

c-Three seedlings hill-1

Irrigation regimes (mm)0 200 400 600 800 1000 1200 1400 1600

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Fig. 8 Simulated water

productivity for the fine rice

cultivar Basmati-385 at variable

plant density and irrigation

regimes. Whiskers represent the

10th and 90th percentile, boxlimits represent the 25th and

75th percentiles, box centralline represents the median, and

outliers represent the minimum

and maximum values for

simulated water productivity

Irrig Sci (2013) 31:491–506 503

123

The monetary return ranged from -483 to 2260 $ ha-1

for the minimum of the lowest and the maximum of the

highest return of simulated scenarios that were simulated.

The monetary return for the 68 scenarios is presented in

Fig. 9. The treatment combination that included two

seedlings hill-1 and an irrigation regime of 1300 mm

showed to be dominant based on the Mean-Gini Domi-

nance Analysis (Buccola and Subaei 1984; Fawcett and

Thornton 1989) among the scenarios that were simulated.

However, the median of these simulated scenarios was very

similar because the difference of the irrigation amount was

only 100 mm among the consecutive scenarios. The min-

imum, 10th, 25th, median, 75th, 90th, and maximum

monetary return ha-1 for this dominant treatment combi-

nation were -12, 297, 844, 1175, 1601, 2027, and 2235 $

ha-1, respectively. The monetary return increased with the

increase in the number of irrigations up to 14 and after that

decreases with the increase in the number of irrigations.

However, the minimum monetary return also remained

negative for first nine irrigation scenarios ranging from

zero (no irrigation) to eight irrigation applications. The

application of system analysis that combines both experi-

mental field research and crop modeling to determine the

most advantageous farming practices for different envi-

ronments has become common. For instance, Sarkar and

Kar (2006) used the CSM-CERES-Rice model to deter-

mine the optimum nitrogen application rates for rice in

India. Arora (2006) used the ORYZA2000 model to

determine the optimum combination of puddling intensities

and irrigation regimes for rice in India.

Conclusions

The model evaluation showed a good agreement for sim-

ulated development and growth parameters for the rice

cultivar Basmati-385 with their respective observed data

obtained from two field experiments having treatment

combinations of irrigation regimes and plant density for

irrigated rice conducted in Faisalabad, Pakistan. This study

illustrated the potential for the application of a decision

support system and crop simulation model to determine

One seedling hill-1

Mon

etar

y re

turn

s ($

ha-1

)

-500

0

500

1000

1500

2000

Two seedlings hill-1

Four seedlings hill-1

Irrigation regimes (mm)0 200 400 600 800 1000 1200 1400 1600

Three seedlings hill-1

-500

0

500

1000

1500

2000

0 200 400 600 800 1000 1200 1400 1600

Fig. 9 Monetary return

($ ha-1) percentile for rice

Basmati-385 at variable plant

density and irrigation regimes.

Whiskers represent 10th and

90th percentile, box plot limitsrepresents the 25th and 75th

percentile, box central linerepresents the median, and

outliers represent the minimum

and maximum values for

monetary returns

504 Irrig Sci (2013) 31:491–506

123

suitable management strategies for rice production in the

Faisalabad district of the Punjab Province of Pakistan.

Although the current Warabandi system for irrigation

management has a lot of limitations with respect to when

and how much water a farmer can apply, this study showed

that there is potential to reduce the number of irrigation

applications, especially toward the end of the growing

season when the crop requires less water. However, further

evaluation of these recommended practices is needed under

field conditions. Based on the results from this study, it can

be concluded that the CSM-CERES-Rice model could

potentially be used as a decision aid to provide alternate

and integrated management options to resource-poor

farmers who grow irrigated rice under semiarid conditions

in different arid or semiarid regions across the world.

Acknowledgments The research was supported in part by the Post

Doctorate Fellowship Program of the Higher Education Commission

(HEC), Islamabad, Pakistan, for the first author of this paper. The

author also appreciates the approval of study leave by the adminis-

tration of Bahauddin Zakariya University (BZU), Multan, Pakistan.

References

Ahmad S, Hussain A, Ali H, Ashfaq A (2005a) Transplanted fine rice

(Oryza sativa L.) productivity as affected by plant density and

irrigation regimes. Intl J Agric Biol 7:445–447

Ahmad S, Hussain A, Ali H, Ahmad A (2005b) Grain yield of

transplanted rice (Oryza sativa L.) as influenced by plant density

and nitrogen fertilization. J Agric Soc Sci 1:212–215

Ahmad S, Zia-ul-Haq M, Ali H, Shad SA, Ahmad A, Maqsood M,

Khan MB, Mehmood S, Hussain A (2008) Water and radiation

use efficiencies of transplanted rice (Oryza sativa L.) at different

plant density and irrigation regimes under semi-arid environ-

ment. Pak J Bot 40:199–209

Ahmad S, Ahmad A, Zia-ul-Haq M, Ali H, Khaliq T, Anjum MA,

Khan MA, Hussain A, Hoogenboom G (2009a) Resources use

efficiency of field grown transplanted rice (Oryza sativa L.)

under irrigated semiarid environment. J Food Agric Environ

7:487–492

Ahmad S, Zia-ul-Haq M, Ali H, Ahmad A, Khan MA, Khaliq T,

Husnain Z, Hussain A, Hoogenboom G (2009b) Morphological

and quality parameters of Oryza sativa L. as affected by

population dynamics, nitrogen fertilization and irrigation

regimes. Pak J Bot 41:1259–1269

Ahmad S, Ahmad A, Soler CMT, Ali H, Zia-ul-Haq M, Anothai J,

Hussain A, Hoogenboom G, Hasanuzzaman M (2011) Applica-

tion of the CSM-CERES-rice model for evaluation of plant

density and nitrogen management of fine transplanted rice for an

irrigated semiarid environment. Precision Agric. doi:10.1007/

s11119-011-9238-1

Arora VK (2006) Application of a rice growth and water balance

model in an irrigated semi-arid subtropical environment. Agric

Water Manage 83:51–57

Banterng P, Hoogenboom G, Patannothai A, Singh P, Wani SP,

Pathak P, Tongpoonpol S, Atichart S, Srihaban P, Bur-

anaviriyakul S, Jintrawet A, Nguyen TC (2010) Application of

the cropping system model (CSM)-CROPGRO-Soybean for

determining optimum management strategies for soybean in

tropical environments. J Agron Crop Sci 196:231–242

Barker R, Dawe D, Tuong TP, Bhuiyan SI, Guerra LC (1999) The

outlook for water resources in the year 2020: challenges for

research on water management in rice production. In: Assess-

ment and orientation towards the 21st century. Proceedings of

19th session of the international rice commissions, Food,

Agriculture Organization, Cairo, Egypt, 7–9 September 1998,

pp 96–109

Borell A, Garside A, Shu FK (1997) Improving efficiency of water for

irrigated rice in a semi-arid tropical environment. Field Crops

Res 52:231–248

Bouman BAM, Tuong TP (2001) Field water management to save

water and increase its productivity in irrigated rice. Agric Water

Manage 49:11–30

Bouman BAM, Shaobing P, Xiaoguang Y, Huaqi W (2002) Aerobic

rice: challenges and opportunities: the price of water-reducing

water requirements in rice production through aerobic system.

In: Paper presented at the international rice congress on

innovation, impact, and livelihood, International Rice Research

Institute, Beijing, September 16–20 (abstract pp 88)

Buccola ST, Subaei A (1984) Mean-Gini analysis, stochastic

efficiency and weak risk aversion. Aust J Agric Econ 28:77–86

Cabangon RJ, Tuong TP, Castillo EG, Bao LX, Lu G, Wang GH, Cui

L, Bouman BAM, Li Y, Chen C, Wang J (2004) Effect of

irrigation method and N-fertilizer management on rice yield,

water productivity and nutrient-use efficiencies in typical

lowland rice conditions in China. Paddy Water Environ

2:195–206

Cheyglinted S, Ranamukhaarachchi SL, Singh G (2001) Assessment

of the CERES-Rice model for rice production in the Central

Plain of Thailand. J Agric Sci 137:289–298

Choudhary MR (2008) A text book of irrigation and drainage:

practices for agriculture, study aid foundation for excellence

(SAFE). Islamabad, Pakistan, pp 32–35

Fawcett RH, Thornton PK (1989) Mean-Gini dominance in decision

analysis. IMA J Math Appl Bus Ind 2:309–317

Feng L, Bouman BAM, Toung TP, Cabangon RJ, Li Y, Lu G, Feng Y

(2007) Exploring options to grow rice using less water in

northern China using a modeling approach I. Field experiments

and model evaluation. Agric Water Manage 88:1–13

Geng S, Zhou Y, Zhang M, Smallwood SK (2001) A sustainable agro-

ecological solution to water shortage in the North China Plain

(Huabei Plain). J Environ Manage 44:345–355

Godwin DC, Singh U (1998) Nitrogen balance and crop response to

nitrogen in upland and lowland cropping systems. In: Tsuji GY,

Hoogenboom G, Thornton PK (eds) Understanding options for

agricultural production. Kluwer Academic, Dordrecht, pp 55–78

GOP (Government of Pakistan) (2007) Economic survey of Pakistan

2006–2007. Finance Division, Economic Advisory Wing,

Finance Division, Government of Pakistan, pp 15–33

GOP (Government of Pakistan) (2008) Economic survey of Pakistan

2007–2008. Economic Advisory Wing, Finance Division, Gov-

ernment of Pakistan, pp 17–36

GOP (Government of Pakistan) (2009) Economic survey of Pakistan

2008–2009. Economic Advisory Wing, Finance Division, Gov-

ernment of Pakistan, pp 17–37

Gungula DT, Kling JG, Togun AO (2003) CERES-maize predictions

of maize phenology under nitrogen-stressed conditions in

Nigeria. Agron J 95:892–899

Hoogenboom G, Jones JW, Wilkens PW, Porter CH, Boote KJ, Hunt

LA, Singh U, Lizaso JL, White JW, Uryasev O, Royce FS,

Ogoshi R, Gijsman AJ, Tsuji GY (2010) Decision support

system for agrotechnology transfer version 4.5 [CD-ROM].

University of Hawaii, Honolulu, Hawaii

IWMI (International Water Management Institute) (2000) IWMI:

global water scarcity study. International Water Management

Institute Colombo, Sri Lanka

Irrig Sci (2013) 31:491–506 505

123

Jintrawet A (1995) A decision support system for rapid assessment of

lowland rice-based cropping alternatives in Thailand. Agric Syst

47:245–258

Jones JW, Hoogenboom G, Porter CH, Boote KJ, Batchelor WD,

Hunt LA, Wilkens PW, Singh U, Gijsman AJ, Ritchie JT (2003)

The DSSAT cropping system model. Eur J Agron 18:235–265

Kato Y, Okami M, Katsura K (2009) Yield potential and water use

efficiency of aerobic rice (Oryza sativa L.) in Japan. Field Crops

Res 113:328–334

Kropff MJ, Bouma J, Jones JW (2001) Systems approaches for the

design of sustainable agro-ecosystems. Agric Syst 70:369–393

Kumar R, Sharma HL (2004) Simulation and validation of CERES-

rice (DSSAT) model in north-western Himalayas. Indian J Agric

Sci 74:133–137

Kunimitsu Y (2009) Measuring the implicit value of paddy irrigation

water: application of RPML model to the contingent choice

experiment data in Japan. Paddy Water Environ 7:177–185

Li Y (2001) Research and practice of water saving irrigation for rice

in China. In: Barker R, Loeve R, Li Y, Tuong TP (eds)

Proceedings of the international workshop on water-saving

irrigation for rice, March 23–25, Wuhan, China. International

Water Management Institute, Colombo, Sri Lanka, pp 135–144

Lin S, Dittert K, Sattelmacher B (2002) The Ground Cover Rice

Production Systems (GCRPS)—a successful new approach to

save water and increase nitrogen fertilizer efficiency? In:

Bouman BAM, Hengsdijk H, Hardy B, Bindraban PS, Tuong

TP, Ladha JK (eds) Water-wise rice production, 8–11 April

2002, LosBanos, Philippines. International Rice Research Insti-

tute, Los Banos, Philippines, pp 187–196

Normile D (2008) Reinventing rice to feed the world. Science

321:330–333

Peng S, Bouman B, Visperas RM, Castaneda A, Nie L, Park H (2006)

Comparison between aerobic and flooded rice in the tropics:

agronomic performance in an eight-season experiment. Field

Crops Res 96:252–259

Rashid MH, Alam MM, Khan MAH, Ladha JK (2009) Productivity

and resource use of direct-(drum)-seeded and transplanted rice in

puddled soils in rice–rice and rice–wheat ecosystems. Field

Crops Res 113:274–281

Ritchie JT (1998) Soil water balance and plant stress. In: Tsuji GY,

Hoogenboom G, Thornton PK (eds) Understanding options for

agricultural production. Kluwer Academic, Dordrecht, pp 41–54

Ritchie JT, Singh U, Godwin DC, Bowen WT (1998) Cereal growth,

development and yield. In: Tsuji GY, Hoogenboom G, Thornton

PK (eds) Understanding options for agricultural production.

Kluwer Academic, Dordrecht, pp 79–98

Ritchie JT, Porter CH, Judge J, Jones JW, Suleiman AA (2009)

Extension of an existing model for soil water evaporation and

redistribution under high water content conditions. Soil Sci Soc

Am J 73:792–801

Sarkar R, Kar S (2006) Evaluation of management strategies for

sustainable rice-wheat cropping system, using DSSAT seasonal

analysis. J Agric Sci 144:421–434

Soler CMT, Hoogenboom G, Sentelhas PC, Duarte AP (2007a)

Growth analysis of maize grown off-season in a subtropical

environment under rainfed and irrigated conditions. J Agron

Crop Sci 193:247–261

Soler CMT, Sentelhas PC, Hoogenboom G (2007b) Application of the

CSM-CERES-Maize model for planting date evaluation and

yield forecasting for maize grown off-season in a subtropical

environment. Eur J Agron 27:165–177

Soundharajan B, Sudheer KP (2009) Deficit irrigation management

for rice using crop growth simulation model in an optimization

framework. Paddy Water Environ 7:135–149

St’astna M, Trnka M, Kren J, Dubrovsky M, Zalud Z (2002)

Evaluation of the CERES models in different production regions

of the Czech Republic. Rostlina Vyroba 48:125–132

Tabbal DF, Bouman BAM, Bhuiyan SI, Sibayan EB, Sattar MA

(2002) On-farm strategies for reducing water input in irrigated

rice; case studies in the Philippines. Agric Water Manage

56:93–112

Thornton PK, Hoogenboom G (1994) A computer program to analyze

single-season crop model outputs. Agron J 86:860–868

Thornton PK, Hoogenboom G, Wilkens PW, Bowen WT (1995) A

computer program to analyze multiple-season crop model

outputs. Agron J 87:131–136

Thornton PK, Hoogenboom G, Wilkens PW, Jones JW (1998)

Seasonal analysis. In: Tsuji GY, Uehara G, Balas S (eds) DSSAT

version 3, vol 3–2. University of Hawaii, Honolulu, pp 1–65

Timsina J, Connor DJ (2001) Productivity and management of rice-

wheat cropping systems: issues and challenges. Field Crops Res

69:93–132

Timsina J, Humphreys E (2006) Performance of CERES-rice and

CERES-wheat models in rice-wheat systems: a review. Agric

Sys 90:5–31

Tsuji GY, Jones JW, Hoogenboom G, Hunt LA, Thornton PK (1994)

Introduction. In: Tsuji G, Uehara Y, Balas GS (eds) DSSAT v3,

Decision support system for agrotechnology transfer, vol 1.

University of Hawaii, Honolulu, p 284

Tsuji GY, Hoogenboom G, Thornton PK (1998) Understanding

options for agricultural production. Systems approaches for

sustainable agricultural development. Kluwer Academic Pub-

lishers, Dordrecht

Tuong TP, Bouman BAM (2003) Rice production in water-scarce

environments. In: Proceedings of water productivity workshop,

12–14 November 2001, Colombo, Sri Lanka. International

Water Management Institute, Colombo, Sri Lanka

Tuong TP, Bouman BAM, Mortimer M (2005) More rice, less water-

integrated approaches for increasing water productivity in

irrigated rice-based systems in Asia. Plant Prod Sci 8:231–241

Wallach D, Goffinet B (1987) Mean squared error of prediction in

models for studying ecological and agronomic systems. Biomet-

rics 43:561–573

Willmott CJ (1982) Some comments on the evaluation of model

performance. Bull Am Meteorol Soc 63:1309–1313

Willmott CJ, Akleson GS, Davis RE, Feddema JJ, Klink KM, Legates

DR, Odonnell J, Rowe CM (1985) Statistics for the evaluation

and comparison of models. J Geophys Res 90:8995–9005

Yang X, Bouman BAM, Wang H, Wang Z, Zhao J, Chen B (2005)

Performance of temperate aerobic rice under different water

regimes in North China. Agric Water Manage 74:107–122

Zeigler RS, Puckridge DW (1995) Improving sustainable productivity

in rice based rainfed lowland systems of South and South-East

Asia. GeoJournal 35:307–324

Zwart SJ, Bastiaanssen WGM (2004) Review of measured crop water

productivity values for irrigated wheat, rice, cotton and maize.

Agric Water Manage 69:115–133

506 Irrig Sci (2013) 31:491–506

123