Embed Size (px)
Citation preview
TI
Ba
b
c
a
ARRA
KVCSWPY
1
i2t(tlr
((
0d
Field Crops Research 124 (2011) 1–13
Contents lists available at ScienceDirect
Field Crops Research
journa l homepage: www.e lsev ier .com/ locate / fc r
he effects of mulch and irrigation management on wheat in Punjab,ndia—Evaluation of the APSIM model
alwinder-Singha,c, D.S. Gaydonb, E. Humphreysc,∗, P.L. Eberbacha
Charles Sturt University, Locked Bag 588, Wagga Wagga, NSW 2678, AustraliaCSIRO Ecosystem Sciences Dutton Park Q 4102, AustraliaInternational Rice Research Institute (IRRI) College, Los Banos, 4031 Laguna, Philippines
r t i c l e i n f o
rticle history:eceived 8 December 2010eceived in revised form 21 April 2011ccepted 28 April 2011
eywords:alidationalibrationoil evaporationater productivity
henologyield
a b s t r a c t
With increasing interest in retaining crop residues on the soil surface, there is a need to evaluate theirshort- and long-term effects on crop yield and water and fertilizer requirements. Therefore, research onthe interactions between residue and irrigation management on wheat crop performance and water usewas initiated, using the dual approach of field experiments and crop modelling. This paper presents theresults of a comprehensive evaluation of the APSIM model for its ability to simulate the effects of mulchand water management, and their interactions, for wheat in Punjab, India. The model was evaluated forits ability to predict crop development, grain yield, biomass production over time, soil water dynamics,daily soil evaporation (Es), total evapotranspiration (ET) and water productivity (WPET kg ha−1 mm−1),using two years of data from field experiments at Ludhiana, Punjab. The model predicted grain yieldadequately, with coefficients of determination (r2) of 0.91 and 0.81 with and without mulch, respec-tively, and prediction of total biomass was even better, with r2 of 0.99 and 0.92. The correspondingabsolute RMSE values were 433 and 550 kg ha−1 for grain yield (means 4100 and 3800 kg ha−1) and 300and 800 kg ha−1 for total biomass (means 10,200 and 9300 kg ha−1). However, grain yield was underpre-dicted (by 600–1000 kg ha−1) in treatments where the crop was subjected to water deficit stress, eventhough simulation of soil water dynamics, and the effect mulch on soil water content, was generally verygood. The model accurately predicted total crop seasonal evaporation and the effect of mulch; however,daily Es was poorly simulated. APSIM does not attempt to capture the soil temperature driven effect of
mulch on crop phenology. The evaluation shows that APSIM is suitable to use for wheat under the con-ditions of north-west India. However, additional model processes that capture the effects of mulch oncrop development and growth, as driven by soil temperature, are needed to help design intensive crop-ping systems to optimise land and water productivity. The ability to better simulate crop performanceunder conditions of water deficit is also needed to help determine irrigation management strategies thatwhil
minimise irrigation input. Introduction
Irrigated wheat is grown in rotation with rice on 2.6 Mha in thentensive rice–wheat system of Punjab in north-west India (GOP,006). However, the sustainability of the rice–wheat system ishreatened by declining soil fertility and groundwater depletionHumphreys et al., 2010; Ladha et al., 2007). The rice–wheat sys-
em also causes serious air pollution as combine harvesting of riceeaves a mixture of loose and anchored residues in the field whichequire burning to enable timely wheat establishment. Therefore,∗ Corresponding author. Tel.: +63 2 580 5600 9x 2342; fax: +63 2 580 5699.E-mail addresses: [email protected] ( Balwinder-Singh), [email protected]
D.S. Gaydon), [email protected] (E. Humphreys), [email protected]. Eberbach).
378-4290/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.fcr.2011.04.016
e maintaining yield.© 2011 Elsevier B.V. All rights reserved.
there has been increasing pressure to find ways of establishing thewheat crop in the rice residues. This has led to the development ofnew sowing machinery, the Happy Seeder (Sidhu et al.,2007, 2008),which enables direct drilling of wheat with full rice residue reten-tion. In addition to reducing air pollution as a result of avoidance ofburning, surface retention of the rice residues offers many potentialbenefits including improved soil physical, chemical and biologi-cal properties, and reduced soil evaporation (Es) (Yadvinder-Singhet al., 2005). However, the effect of surface residue retention on cropperformance and water use is dependent on soil type, weather con-ditions and amount of residue. Therefore, the inclusion of surfaceresidue retention in the cropping system requires greater knowl-
edge of its effects on crop performance and nutrient and waterdynamics on a long-term basis. Recently, there have been severalfield studies which focus on the effect of mulch on wheat yield innorth-west India (Balwinder-Singh et al., 2011; Chakraborty et al.,
2 d Crop
22sbm
sswtoipiJleiPuwas
a(1tmFraomsttawTAtap
2
2
mIWmcdT11
m(mKt
Balwinder-Singh et al. / Fiel
008; Sharma et al., 2008; Sidhu et al., 2007; Yadvinder-Singh et al.,008). However, the results of these studies are season- and site-pecific, short-term, and provide few insights into the mechanismsy which mulch affects yield and water availability and require-ent.Cropping system models are useful tools for studying the
oil–crop–atmosphere system as affected by management (e.g.owing date, fertilizer application, irrigation scheduling) andeather variability. They can also be used to extrapolate results
o other sites and climates over space and/or time. To date, vari-us crop models have been successfully evaluated and applied tondividual crops including rice and wheat, and occasionally crop-ing system models have been applied to simulate crop sequences,
n north-west India (e.g. Arora et al., 2007; Jalota and Arora, 2002;alota et al., 2006; Timsina et al., 2008). However, there are no pub-ished data on the performance of crop models in simulating theffects of surface residue retention on crop performance, nor itsmplications for irrigation management. The APSIM (Agriculturalroduction System Simulator) cropping system model allows sim-lation of residue and irrigation management on crop performance,ater dynamics, soil organic carbon (C) and soil nitrogen (N) ondaily basis (Keating et al., 2003) for individual seasons or crop
equences over one to many years.APSIM has been validated and successfully used for wheat over
broad range of soils and climates in various part of the worlde.g. Asseng et al., 1998, 2000; Keating et al., 1995, Meinke et al.,998; Yunusa et al., 2004). However, it has not been tested inhe rice–wheat system of the Indo-Gangetic Plains (IGP), India’s
ajor wheat growing region. Moreover, the APSIM module SUR-ACE ORGANIC MATTER, which simulates the dynamics of surfaceesidues and their influence on the components of the water bal-nce, has not been evaluated for its ability to simulate the effectsf surface residue retention on detailed dynamics of Es. Before aodel can be used to help determine optimum crop and cropping
ystem management, it must be parameterised and validated forhe environments and management practices of interest. Thereforehis paper presents the results of the parameterisation and evalu-tion of APSIM for wheat, established in bare soil or rice residues,ith a range of irrigation scheduling treatments, in Punjab, India.
he objective was to determine the level of confidence with whichPSIM could be used for subsequent simulation studies to examine
he effects of surface residue retention and irrigation management,nd their interactions, on crop performance, water use and waterroductivity.
. Materials and methods
.1. APSIM model (v. 5.1)
APSIM is a simulation modelling framework that enables sub-odels to be linked to simulate agricultural system performance.
n simulating wheat crops, the four modules used are WHEAT, SOIL-AT2, SOILN2 and SURFACE ORGANIC MATTER. The WHEAT cropodule simulates the development, growth, water and N uptake,
rop N concentration, stresses (water deficit, N deficit, aerationeficit) and response of the crop to the stresses (Keating et al., 2001).he WHEAT module is based on CERES Wheat (Jones and Kiniry,986; Ritchie et al., 1985) but with modifications (Asseng et al.,998; Probert et al., 1995; Wang et al., 2003; Huth, unpublished).
The soil water module (SOILWAT2) is a cascading water balanceodel based on the water balance models in CERES and PERFECT
Probert et al., 1998; Littleboy et al., 1992). Enhancements to theodule over these models are described in Asseng et al. (1998) and
eating et al. (2003). Soil evaporation in SOILWAT2 is assumed toake place in two stages, following the approach of Ritchie (1972).
s Research 124 (2011) 1–13
These two stages are described through the use of two parameters:U and cona. The parameter U represents the amount of cumula-tive Es before the rate of soil water supply at the surface decreasesbelow atmospheric demand. The rate of Es during the second stageis specified by the parameter cona as a function of the square rootof time (t) since the end of first stage evaporation as follows:
Es = cona × t1/2.
The value of cona depends on soil hydraulic properties andpotential evapotranspiration (ETo) (Prihar et al., 1996; Ritchie,1972). Potential evapotranspiration is calculated daily using anequilibrium evaporation concept as modified by Priestley andTaylor (1972).
The APSIM SURFACE ORGANIC MATTER module (formerlyAPSIM-RESIDUE) was developed by Probert et al. (1995, 1998) andis described in detail by Thorburn et al. (2001). The effect of cropresidues on runoff is accounted for by modifying the USDA curvenumber using the relationships from Glanville et al. (1984) whichdemonstrated that curve number decreases by one unit for every4% of crop residue cover up to a maximum of 80% cover. The effectof crop residues on the potential drying rate of soil water is alsocalculated, based on the results of (Adams et al., 1976). SOILN2 isbased on the CERES model (Ritchie et al., 1985), with modifications(Probert et al., 1995, 1998). Mineral N is considered as NO3
−, NH4+
and urea. The transformation processes of mineralization, immo-bilization, denitrification and urea hydrolysis are all simulated inSOILN2.
2.2. Data sets for model calibration and evaluation
2.2.1. Data for model inputsReplicated field experiments comparing wheat direct-drilled
into rice residues and bare soil were conducted on a clay loam soilin Punjab, India, in 2006–2007 and 2007–2008. There were twomulching treatments (with and without mulch) and six irrigationscheduling treatments:
• I1—irrigation applied when soil water tension increased to 50 kPaat 15–20 cm soil depth for the first irrigation, and at 35–40 cm forsubsequent irrigations;
• I2 (control)—irrigation around the time of crown root initiation(36 d after sowing (DAS) in 2006, 27 DAS in 2007) and thereafterwhen the ratio of the amount of irrigation water (IW) appliedat the previous irrigation to cumulative pan evaporation minusrain (CPE-rain) increased to 0.9, i.e. IW/(CPE-rain) = 0.9 (recom-mended practice, Prihar et al., 1974);
• I3—same as I2 minus the irrigation at crown root initiation (CRI);• I4—one irrigation at CRI then irrigation when IW/(CPE-rain) = 0.6;• I5—one irrigation only, at CRI;• I6—as for I2 minus the last irrigation.
Seventy-five (75) mm of water were applied at each irrigationfor all treatments, therefore IW/(CPE-rain) ratios of 0.9 and 0.6 areequivalent to net CPE of 83 and 125 mm, respectively, betweenirrigations. Soil water tension in I1 was measured using tube ten-siometers and a SoilSpec® vacuum gauge. The tensiometers wereinstalled mid-way between the plant rows with the tips at 15–20and 35–40 cm. In 2006–2007, irrigation was applied 36 DAS to alltreatments, and because of the well-distributed rainfall, there wasno further irrigation of any treatment except for the non-mulchedtreatment with irrigations scheduled by tensiometer (I1) which
received one irrigation 76 DAS. Therefore only two irrigation treat-ments could be implemented in 2006–2007—I1 and I2. However, in2007–2008, rainfall was low and poorly distributed and all six treat-ments were implemented; each irrigation treatment was different
d Crops Research 124 (2011) 1–13 3
ie
r(pjcS
tWl1rtwdmt
mdswih
R
wtcscsc
y = 0.994xR² = 0.856
0
5
10
15
20
25
30
302520151050
Cal
cula
ted
radi
atio
n (M
J m
-2 d
ay-1
)
Pyranometer (MJ m-2 day-1)
−2 −1
TS
LDSB
TC
Balwinder-Singh et al. / Fiel
n amount (75–225 mm) or time of application (Balwinder-Singht al., 2011).
Wheat (variety PBW343) was sown at 100 kg ha−1, 20 cmow-spacing in early November, with diammonium phosphate23 kg N ha−1 and 26 kg P ha−1). Urea was broadcast at 37 kg N ha−1
rior to sowing, and a further 60 kg N ha−1 (as urea) was broadcastust prior to the time of the irrigation at crown root initiation. Siteonditions, weather, and management are presented in Balwinder-ingh et al. (2010, 2011).
Soil physical properties and initial volumetric soil water con-ent (VWC) (0–180 cm) and mineral N are presented in Table 1.
ater content at the lower limit and drained upper limit of eachayer was determined using pressure plate apparatus at 0.3 and5 bar, respectively. Initial VWC was determined from gravimet-ic water content of soil samples (collected using an auger at theime of sowing) and bulk density. Total mineral N (NO3
− and NH4+)
as determined using 2 M KCl extraction (1:10) followed by steamistillation and titration (Keeney and Nelson, 1982). Daily Es waseasured using mini-lysimeters (Balwinder-Singh et al., 2010) for
he whole crop season.The minimum weather data needed for APSIM are daily maxi-
um and minimum air temperature (◦C), daily rainfall (mm) andaily incoming solar radiation (MJ m−2 day−1). Rainfall was mea-ured at the field site, and temperature and daily sunshine hoursere measured at a weather station about 1 km from the exper-
mental site. Solar radiation was calculated from daily sunshineours using the Angstrom Formula (Sys et al., 1991):
s = Ra(a + b)(n/N)),
here: Rs is total radiation received at the earth’s surface; Ra ishe extra terrestrial radiation at the edge of the atmosphere andalculated for each day from latitude and day of year; n is the sun-
hine hours; N is the day length (hours); a and b are the Angstromoefficients and were calculated from the latitude using equationsuggested by Bandyopadhyay et al. (2008). For Ludhiana, the cal-ulated values of a and b were 0.2176 and 0.4919, respectively.able 1oil physical properties and initial soil moisture at sowing during two cropping seasons.
Soil layer(cm)
Clay (%) LL(cm3 cm−3)
DUL(cm3 cm−3)
BD (g cm−3)
0–15 32.7 0.10 0.31 0.35 1.5015–30 41.2 0.12 0.32 0.39 1.6130–60 45.1 0.11 0.33 0.38 1.4660–90 38.6 0.09 0.31 0.41 1.4890–120 15.8 0.07 0.24 0.40 1.33120–150 4.3 0.05 0.20 0.38 1.39150–180 4.2 0.05 0.20 0.38 1.42
L—volumetric water content at wheat crop lower limit.UL—volumetric water content at drained upper limit.AT—volumetric water content at saturation.D—bulk density.
able 2omparison of simulated and observed dates of phenological stages in treatment I2.
Residue treatment 2006–2007
Simulated
Emergence Non-mulch 13 NovemberMulch 13 November
Flowering Non-mulch 21 FebruaryMulch 21 February
Physiological maturity Non-mulch 14 AprilMulch 14 April
Fig. 1. Calculated solar radiation (MJ m day ) from sunshine hours comparedwith measurements by pyranometer during the 2006–2007 wheat season.
Radiation data were also collected at the site in 2007–2008 usingan Apogee® pyranometer, and the agreement between measuredradiation and that calculated using the Angstrom Formula was good(r2 = 0.86; Fig. 1).
2.2.2. Data for model calibration and evaluationThe details of the field data used for model calibration and eval-
uation are provided in Balwinder-Singh et al. (2010, 2011). The cropdata include the dates of 50% anthesis and physiological maturity,total biomass at eight times throughout the season, grain yield andyield components for all treatment combinations. The water datainclude irrigation time and amount (all treatments). Volumetricsoil water content and ET (calculated from the change in VWC)
were determined approximately twice weekly, and Es was deter-mined daily, in selected treatments (I2 in 2006–2007, and I1 and I2in 2007–2008).Roothospitalityfactor (xf)
Initial soil water(cm−3 cm−3)
Initial soil mineral N(NH4
+ + NO3−)(kg ha−1)
2006–2007 2007–2008 2006–2007 2007–2008
1.0 0.24 0.19 30 350.7 0.27 0.25 22 240.3 0.32 0.30 – –0.2 0.31 0.28 – –0.1 0.11 0.20 – –0.0 0.07 0.11 – –0.0 0.06 0.07 – –
2007–2008
Observed Simulated Observed
13 November 20 November 20 November15 November 20 November 22 November19 February 6 March 3 March25 February 6 March 12 March8 April 15 April 10 April18 April 15 April 16 April
4 Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
(kg
ha-1
)
Days after sowing
(a) I1
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
( ha-
1 )
Days after sowing
(b) I2
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
(kg
ha-1
)
Days after sowing
(c) I1
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
(kg
ha-1
)
Days after sowing
(d) I2
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
(kg
ha-1
)
Days after sowing
(e) I3
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
(kg
ha-1
)
Days after sowing
(f) I4
F ch (dat (g) I5,
2
seayutmtfiip
ig. 2. Simulated (lines) and observed (symbols) above ground biomass under mulreatments during 2006–2007 (a) I1, (b) I2 and 2007–2008, (c) I1, (d) I2, (e) I3, (f) I4,
.3. Model calibration
The well-watered irrigation treatment I2 from 2007–2008 waselected for calibration of wheat variety PBW343 using the fieldxperimental data for the number of days from sowing to 50%nthesis and physiological maturity (12.5% grain moisture), grainield, and total biomass. The genetic coefficients were estimatedsing the best fit method, i.e. by iteratively varying the values ofhe coefficients to produce a close match between simulated and
easured phenology, grain yield and total biomass. The values of
he coefficients for wheat cv. PBW343 were: startgf to mat (grainlling duration in degree days, ◦C) = 750, phyllochron (phyllochronnterval, ◦C) = 95.0, vern sens (sensitivity to vernalisation) = 1.7,hotop sens (photoperiod sensitivity) = 3.8. These values are close
rk lines/filled symbols) and non-mulch (dotted lines/hollow symbols) in irrigation(h) I6 (Vertical bars are standard error of the means of the observed values).
to those reported by Timsina and Humphreys (2006) for wheatcultivars with similar characteristics to PBW343, for use in CERES-Wheat.
The Es coefficients ‘U’ (11 mm) and ‘cona’ (4.0 mm1/2) wereestimated from field measured cumulative Es during 2007–2008(non-mulched treatment) (Fig. 8c), and were close to the valuesused by Arora et al. (2007) in same environment on a sandy loamsoil.
The root hospitality factor (Keating et al., 2003) in eachsoil layer was derived from wheat root distribution field
studies (Yadvinder-Singh et al., 2009) and the values usedby Timsina et al. (2008) in the same region. These val-ues were then adjusted until a good match was achievedbetween the simulated and observed soil water extraction pat-
Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13 5
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
( kg
ha-1
)
Days after sowing
(g) I5
0
2000
4000
6000
8000
10000
12000
0 20 40 60 80 100 120 140 160 180
Bio
mas
s pr
oduc
tion
(kg
ha-1
)
Days after sowing
(h) I6
Fig. 2. (Continued) .
Table 3Absolute deviation of simulated grain yield (kg ha−1) and total biomass (t ha−1).
Grain yield (kg ha−1) Total biomass (t ha−1)
Non-mulch Mulch Non-mulch Mulch
Sim. Meas. Dev. (%) Sim. Meas. Dev. (%) Sim. Meas. Dev. (%) Sim. Meas. Dev. (%)
2006–2007I1 4400 3900 +12.5 4400 4400 0.0 10,700 9300 +15.0 10,400 10,500 −0.1I2 4200 4000 +4.3 4400 4500 −3.0 9700 9200 +5.4 10,400 10,600 −1.9Mean 4273 3945 +8.4 4380 4441 −1.4 10,200 9300 +9.7 10,400 10,600 −1.9
2007–2008I1 3400 3600 −7.3 3800 3850 −0.1 9200 9600 -4.2 9400 9700 −3.0I2 4300 4000 +6.9 4700 4100 +14.9 10,100 9200 +9.8 10,800 10,300 +4.6I3 3700 4200 −13.7 4000 4300 −8.7 8900 9900 −10.1 9300 9800 −5.1I4 3300 3600 −8.4 3800 3700 +3.4 8500 8800 −3.5 9300 9500 −2.2
032
tT
2
aataoasESttasto
R
N
wa
I5 2100 3100 −33.2 2400 3300 −27.I6 3000 3500 −15.5 3450 3900 −12.Mean 3269 3678 −11.1 3684 3855 −5.
ern during the 2006–2007 crop season, and are presented inable 1.
.4. Model evaluation
Simulated model output was compared with observed values forrange of crop parameters from all treatments in the 2006–2007
nd 2007–2008 experiments, and for a range of water parameters inhe two treatments where these were monitored (I2 in 2006–2007,nd I1 and I2 in 2007–2008). These parameters included the datesf 50% emergence, 50% anthesis and physiological maturity, theccumulation of biomass over time, grain yield, VWC of differentoil layers throughout the season, daily and cumulative Es, totalT, and water productivity with respect to ET (WPET) (Balwinder-ingh et al., 2010, 2011). Model performance was assessed usinghe absolute and normalized root mean square error (RMSE), andhe coefficient of determination (r2) of the regression of observedgainst simulated values through the origin. Low RMSE (of theame, or less, order of magnitude as experimental standard devia-ions) and high r2 values indicate good agreement between modelutputs and observed values.
MSE = N−1
(∑n
i=1(Pi − Oi)
2)0.5
( )
ormalized RSME(%) = Absolute RMSEMean of the observed× 100,
here Pi and Oi are predicted and observed values, respectively,nd N is number of observations.
6900 8800 −21.6 7600 8600 −11.68800 9000 −2.2 9500 9800 −3.08700 9200 −5.4 9300 9600 −3.1
3. Results
3.1. Phenology
There was good agreement between simulated and observeddates of emergence, anthesis and physiological maturity for thenon-mulched irrigation treatments each season (there was noobserved difference in phenological dates between irrigation treat-ments) (Table 2). However, the model did not simulate thedevelopmental delay observed in the mulched treatments. With-out mulch, both observed and simulated emergence occurred 7DAS each year, simulated anthesis occurred 2–3 d after observedanthesis, and simulated physiological maturity occurred 5–6 d afterobserved maturity. With mulch, the simulated dates of all develop-mental stages were the same as those without mulch, each season.However, actual emergence (through soil surface) was delayed byabout two days in both seasons with mulch (Balwinder-Singh et al.,2011), as the model does not simulate the effect of mulch on lower-ing soil temperature and the consequent delay in emergence. Mulchalso delayed anthesis by 7 and 5 d in 2007 and 2008, respectively,while physiological maturity was delayed by 10 and 6 d, in the fieldexperiments.
3.2. Biomass
In 2006–2007, there was generally good agreement betweensimulated and observed accumulation of biomass over time. In2007–2008, simulation of early growth was good in all treat-ment combinations (Fig. 2c–h). However, during an unusually cool
6 Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13
Table 4Model evaluation parameters for different model outputs attributes.
Model attribute Residue treatment Number of data points r2 RMSE Normalized RMSE (%)
Grain yield (kg ha−1) Mulch 8 0.91 443 12.4Non-mulch 8 0.86 550 16.5
Time course of Biomass (kg ha−1) Mulch 56 0.95 800 23.4Non-mulch 56 0.95 600 17.9
Biomass at Maturity (kg ha−1) Mulch 8 0.99 300 3.6Non-mulch 8 0.92 800 10.8
Daily Es (mm) Mulch 307 0.06 0.65 90.0Non-mulch 307 0.13 0.71 78.5
Total water use/(ET, mm) Mulch 8 0.79 50.8 14.2Non-mulch 8 0.93 32.1 9.0
WPET Mulch 8 0.37 1.6 14.1Non-mulch 8 0.30 1.3 12.7
Volumetric water content (both residue treatments)0–15 cm 200 0.43 0.05 25.2
0.47 0.04 18.40.72 0.04 15.20.88 0.03 13.4
ptieetbtimnTwttmamb
3
ysm2wyma(tptafip
3
i
Observed grain yield (kg ha-1)5000400030002000
Sim
ulat
ed g
rain
yie
ld (k
g ha
-1)
2000
3000
4000
5000
Observed biomass (kg ha-1)1200011000100009000800070006000
Sim
ulat
ed b
iom
ass
(kg
ha-1
)
6000
7000
8000
9000
10000
11000
12000
15–30 cm 20030–60 cm 20060–90 cm 152
eriod in January 2008 (Balwinder-Singh et al., 2011), the simula-ions overestimated biomass, more so in the mulched treatments,n which there was very little growth between mid-January toarly February. Both the mulched and non-mulched crops recov-red from this effect as the temperature rose, and by maturityhere was good agreement between simulated and observed totaliomass in all treatment combinations except for the non-mulchedreatment which received only one post sowing irrigation (I5) dur-ng this relatively dry year (Fig. 2g). Total biomass predictions at
aturity were generally close to observed values, except for theon-mulched I5 in which it was underestimated by 21.6% (Table 3).he model’s performance in terms of the evaluation parametersas good. The coefficient of determination (r2) was high for both
reatments (r2 = 0.95) (Fig. 3; Table 4). The absolute RMSE for theime course of biomass for the mulched and non-mulched treat-
ents was 800 and 600 kg ha−1, and normalized RMSE was 23.4nd 17.9% for same treatments, respectively. The higher values forulch than non-mulch indicate reduced precision in simulating
iomass production with mulch.
.3. Grain yield
There was good agreement between simulated and observedield for treatments that did not suffer from severe water deficittress during the grain filling period (I1 and I2 with and withoutulch in 2006–2007; I1, I2, I3 and I4 with and without mulch in
007–2008) (Table 3). In these treatments, predicted yields wereithin 0–13% of observed yields. The deviation from observed grain
ield was greatest (underestimated by 27 and 33% with and withoutulch, respectively) in I5 in 2007–2008, which was not irrigated
fter CRI. Grain yield predictions were good in the well-irrigatedhigher yielding) treatments, but yields were underpredicted inhe lower yielding, poorly irrigated treatments (Fig. 3). The modelredicted higher yields with mulching than without mulching forhe same irrigation management, consistent with observations. Thebsolute RMSE, normalized RMSE and r2 show that the model per-ormed well in predicting grain yield with and without mulch,n contrasting seasonal conditions, but that the predictions wasoorer with mulch (Table 4).
.4. Soil water
APSIM captured all the main features of soil water dynam-cs throughout the soil profile in both years, with and without
Fig. 3. Simulated and observed grain yield and total biomass during both yearsunder all residue and irrigation treatments.
mulch, for the well-irrigated crops (treatments I1 and I2) each year(Figs. 4–6). There was consistently slight over-prediction of VWC inthe top two soil layers (0–15, 15–30 cm) after sowing and after thefirst irrigation in all treatments and years, especially in 2006–2007
without mulch. The difference was generally greater in the top layer(0–15 cm). However, after anthesis, simulated water content in the0–15 cm soil layer was usually close to observed values. In deepersoil layers there was always very good agreement between simu-
Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13 7
Table 5Absolute deviation of simulated crop water use and water productivity from field observed data.
Water use (mm) WPET (kg ha−1 mm−1)
Non-mulch Mulch Non-mulch Mulch
Sim. Meas. Dev. (%) Sim. Meas. Dev. (%) Sim. Meas. Dev. (%) Sim. Meas. Dev. (%)
2006–07I1 341 367 −7.1 312 360 −13.3 12.8 10.6 +20.5 14.0 12.1 +15.7I2 318 345 −7.8 312 341 −8.5 13.1 11.6 +12.9 14.0 13.2 +6.0Mean 330. 356 −7.3 312 351 +11.1 13.0 11.1 +17.1 14.0 12.7 +10.22007–08I1 356 384 −7.3 314 363 −13.4 9.5 9.4 +1.0 12.1 10.6 +14.1I2 384 404 −5.0 366 400 −8.5 11.1 9.9 +12.1 12.8 10.2 +25.4I3 311 350 −11.1 301 369 −18.4 11.7 12.0 −2.5 13.2 11.8 +11.8I4 321 354 −9.3 312 362 −13.8 10.3 10.2 −1.0 12.2 10.2 +19.6I5 244 290 −15.8 236 306 −22.8 8.6 10.8 −20.4 10.1 10.7 −5.7I6 322 353 −8.8 313 357 −12.3 9.2 10.0 −8.0 11.0 11.0 0.0Mean 323 356 −9.2 307 360 −14.7 10.0 10.4 −3.8 11.9 10.7 11.2
0-15cm
VWC
(cm
3 cm
− 3)
0.05
0.15
0.25
0.350-15cm
30-60cm30-60cm
0.05
0.15
0.25
0.35
60-90cm
1701501301109070503010
0.05
0.15
0.25
0.35
15-30cm
Non-mulch Mulch
15-30cm
0.05
0.15
0.25
0.35
Days after sowing
60-90cm
1701501301109070503010
Fig. 4. Comparison of simulated and measured volumetric soil water content (cm3 cm−3) in soil profile (0–90 cm) with and without mulching during the 2006–2007 wheatseason in irrigation treatment I2 (Vertical bars are standard error of the means of the observed values).
8 Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13
MulchNon-mulch0-15cm
0.05
0.15
0.25
0.35
30-60cm
0.05
0.15
0.25
0.35
0-15cm
30-60cm
60-90cm
Days after sowing1701501301109070503010
15-30cm
VWC
(cm
3 cm
− 3)
0.05
0.15
0.25
0.35
0.4515-30cm
60-90cm
1701501301109070503010
0.05
0.15
0.25
0.35
F cm−3)s e obs
ltt2a
e(elgysfitsapl
ig. 5. Comparison of simulated and measured volumetric soil water content (cm3
eason in irrigation treatment I1 (Vertical bars are standard error of the means of th
ated and observed VWC. The r2 increased from 0.43 at 0–15 cmo 0.88 at 60–90 cm soil depth. For the surface layer (0–15 cm)he absolute and normalized RMSE values were 0.05 cm3 cm−3 and5.2%, respectively, which decreased to 0.03 cm3 cm−3 and 13.4%t 60–90 cm (Table 4).
The model predicted higher soil water content in the upper lay-rs of the mulched treatments than the non-mulched treatmentse.g. Fig. 7), consistent with the observed data (Balwinder-Singht al., 2010, 2011). The simulations showed that this was due toower Es in the mulched treatments. Simulated cumulative Es wasenerally close to observed Es, with and without mulch, in I2 eachear (Fig. 8a–d). The model tended to overestimate daily Es afterowing in the mulched treatment each year, and more so after therst irrigation in both mulching treatments. Conversely, the modelended to underestimate daily Es for 2–3 weeks prior to anthe-
is. The net result was very good agreement between predictednd observed total Es both with and without mulch. However,rediction of daily Es was relatively poor, with high values of abso-ute and normalized RMSE for both mulching treatments (Table 4),
in soil profile (0–90 cm) with and without mulching during the 2007–2008 wheaterved values).
although the predicted values fell within the confidence interval ofthe observed values. Mulch reduced simulated total Es by 22 and37 mm in 2006–2007 and 2007–2008, compared to the observedreductions of 34 and 40 mm in respective years.
The model predicted no drainage beyond 90 cm, consistent withthe very small changes in soil matric potential and volumetric watercontent in the deeper soil layers (>120 cm) during the two years inI1 and I2 (data not presented).
3.5. Crop water use (ET) and water productivity
The model was reasonably good at simulating total crop wateruse (ET) (Table 4). Simulated ET was slightly lower than observed ETin all treatment combinations in both years, with higher differencesunder mulch, and with the biggest deviation of 22.8% in the treat-
ment with greatest water deficit (I5 in 2007–2008) (Table 5). Thedeviation ranged from −8.5 to −22.8% with mulch and from −7.1 to−15.8% without mulch. Greater underprediction with mulch maybe due to the fact that the water saved from loss as Es was not fully
Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13 9
30-60cm
0.05
0.15
0.25
0.3530-60cm
60-90cm
10 30 50 70 90 110 130 150 170
15-30cm15-30cm
VWC
(cm
3 cm
−3)
0.05
0.15
0.25
0.35
60-90cm
Days after sowing10 30 50 70 90 110 130 150 170
0.05
0.15
0.25
0.35
0-15cmMulch
0-15cmNon-mulch
0.05
0.15
0.25
0.35
F cm−3)s e obs
dih
(lwwlsoe
4
pws
ig. 6. Comparison of simulated and measured volumetric soil water content (cm3
eason in irrigation treatment I2 (Vertical bars are standard error of the means of th
iverted to transpiration by the model, resulting in lower ET. In allrrigation treatments, the simulated reduction in Es with mulch wasigher than the simulated increase in transpiration with mulch.
Wheat water productivity was calculated based on grain yielddry) and ET. There was generally good agreement between simu-ated and observed WPET, and simulated WPET was always higher
ith mulch, as in I1, I2 and I6 the field experiments (Table 5). WPETas underpredicted by 20.4% in the non-mulched I5, reflecting the
arge underprediction of grain yield (by 32.0%) and the lower butubstantial underprediction of ET (by 15.8%). APSIM’s predictionsf WPET were acceptable, and with similar values of the modelvaluation parameters with and without mulch (Table 4).
. Discussion
The overall performance of APSIM in simulating wheat croperformance and water dynamics, for mulched and non-mulchedheat in north-west India, was very good. However, there were
ome conditions under which the model did not perform well in the
in the soil profile (0–90 cm) with and without mulch during the 2007–2008 wheaterved values).
simulation of some variables, especially in the mulched treatments.Grain yield predications were good for all treatments other thanthose with the greatest water deficit stress, and the magnitude ofthese evaluation parameters was similar to the with CERES-Wheatin the same region (Timsina and Humphreys, 2006).
4.1. Simulation of phenology
The phenology predictions were close to the observed datesin the non-mulched treatments, and the errors were withinrange observed in other simulation studies using CERES-Wheatin the same region (Arora et al., 2007; Hundal and PrabhjyotKaur, 1997). However, the model did not predict the delay indevelopment in the presence of mulch observed in the field byBalwinder-Singh et al. (2011) and Sidhu et al. (2007). Like most
crop models, APSIM-Wheat calculates phenological stages basedon thermal time (cumulative degree days). Thermal time is cal-culated using air temperature and does not reflect that fact thatsurface mulch alters soil temperature in the seed zone, and that
10 Balwinder-Singh et al. / Field Crop
0-15 cm
0.075
0.125
0.175
0.225
0.275
0.325
0.375
Non-mulchMulch
15-30 cm
Days after sowing1701501301109070503010
VWC
(cm
3 cm−3
)
0.125
0.175
0.225
0.275
0.325
0.375
Non-mulchMulch
Fig. 7. Comparison of simulated volumetric water content (cm3 cm−3) under mulchand non-mulched wheat in the top two soil layers in irrigation treatment I2 duringt
t(eodtptiRr1aPT1
4
taarboi
he 2007–2008 wheat season.
his affects the rates of germination, development and growthChaudhary and Chopra, 1983; Kaspar and Bland, 1992; Tripathit al., 1985). Chaudhary and Chopra (1983) found that in sub-ptimal temperature conditions, the decrease in soil temperatureue to rice straw mulch resulted in slower growth and lowerotal biomass at maturity. In wheat, an increase in soil tem-erature towards optimum (mean temperature of 19◦C) aroundhe crown root nodes (in the top 3–5 cm soil layer) significantlyncreases total dry matter production (Boatwright et al., 1976).oot (soil) temperature has been found to control the rates ofoot differentiation, lateral root growth (Miyasaka and Grunes,990), shoot growth, senescence of aerial parts (Kuroyanagind Paulsen, 1988), grain starch accumulation (Guedira andaulsen, 1988), leaf appearance and leaf extension (Hay andunnicliffe, 1982), and photosynthetic activity (Udomprasert et al.,995).
.2. Effect of low temperature on simulation of biomass
The model was also unable to simulate the effect of lowemperature on biomass production during tillering in 2008,
nd the subsequent recovery of the crop (higher observed thanctual growth rate after temperature increased). However, the netesult was good agreement between simulated and observed totaliomass at maturity in all treatments (except for underestimationf biomass in treatments with the greatest water deficit stress dur-ng grain filling, see below).s Research 124 (2011) 1–13
4.3. Simulation of biomass and yield under mild to moderatewater deficit stress
The model generally captured the effect of the irrigation treat-ments on yield, except for the treatment with no irrigation at crownroot initiation (I3). In this treatment measured yields were similarto or higher than yields of the recommended irrigation practice(I2), with and without mulch, whereas the simulated yields of I3were much lower than the simulated yields of I2. Biomass pro-duction during the grain filling period, and grain yield, were alsogreatly underestimated by the model in the treatment with thegreatest water deficit stress from around heading to maturity (I5).Grain yield underestimation in water stressed treatments was alsoreported by Asseng et al. (1998) and may be due to insufficientremobilization of stored pre-anthesis carbohydrates to the grain bythe model. They suggested that model prediction may be improvedby including routines which transfer additional carbohydrates onthe basis of a fraction of stem weight into the grain when cropsexperience severe drought conditions.
4.4. Simulation of the effect of mulching
The model predicted a positive effect of mulch on wheat grainyield consistent with the observed data, and also with the find-ings of others in same environment (Chakraborty et al., 2008;Sidhu et al., 2007; Yadvinder-Singh et al., 2008). The higher simu-lated grain yield with mulch was associated with higher simulatedtranspiration (by 18 and 22 mm in 2006–2007 and 2007–2008,respectively), consistent with the findings of Balwinder-Singh et al.(2010). It was also associated with lower post anthesis water stressunder mulch, evident from the lower values of the water stressindices swdef photo and swdef expan (water stress for photo-synthesis and leaf expansion, respectively). For example, duringthe period from flowering to the end of grain filling, the waterstress indices were higher in non-mulched wheat (swdef photo0.18, swdef expan 0.24) than with mulch (swdef photo 0.07,swdef photo 0.12), where a value of 1 = high stress and 0 = no stress.
4.5. Simulation of soil water dynamics
The SOILWAT2 module performed well in predicting soil waterdynamics in the mulching and irrigation treatments where VWCwas monitored. The overestimation of VWC in the topsoil, espe-cially in non-mulched wheat in 2006–2007, may be due todifferences in simulated and actual daily Es rates. The origin ofthese differences requires further investigation, but is possibly dueto uncertainties in APSIM inputs such as saturated flow parameters(mwcon) which govern downward water movement under condi-tions above saturation (such as immediately after irrigation). It mayalso be due to sub-optimal process description. The inclusion ofuser defined specific root hospitality factors in the original APSIMapproach (Keating et al., 2001) resulted in good predictions of thetime course and final depth of the simulated root system and alsoin good prediction of soil water content in deeper layers where soilwater changes are controlled by root growth. This was also the casein the present study.
Simulated total Es was close to observed values in both residuetreatments in the irrigation treatment (I2) where Es was mea-sured. The model over-predicted the second stage Es after sowingand after the first irrigation. The reverse trend was observed laterin the season, with underprediction of Es. As a result, the simu-lated and observed cumulative Es converged. Over-prediction of
Es during the initial period of second stage evaporation was alsoobserved by Prihar et al. (1996). Using Ritchie’s approach, theyobtained different slopes for second stage Es for different levelsof evaporative demand for the same soil; therefore, they used
Balwinder-Singh et al. / Field Crops Research 124 (2011) 1–13 11
(a) 2006-07-Non-mulch
10
30
50
70
90
110
130
150
170
SimulatedMeasured
(b) 2006-07-Mulch
(c) 2007-08-Non-mulch
s af
1701501301109070503010
Cum
ulat
ive
soil
evap
orat
ion
(mm
)
10
30
50
70
90
110
130
150
170 (d) 2007-08- Mulch
1701501301109070503010
F aporad
doaocWuelatiioel
rtaatamosmupe(
Day
ig. 8. Simulated (filled symbols) and observed (hollow symbols) cumulative soil ev) (Vertical bars are standard error of the means of the observed values).
ifferent second stage constant values for different levels of evap-rative demand. Aydin et al. (2005) also reported that Ritchie’spproach is unable to predict daily Es accurately, but is capablef providing good estimates of cumulative Es over hydrologi-ally significant periods, i.e. weeks and months, as in our study.
allace et al. (1993) also observed that there was fairly highncertainty in prediction of individual daily values of Es, but thatstimation of cumulative Es was good. In studies to analyse theong-term impacts of management practices on water savingsnd water productivity by moderating Es, accurate simulation ofotal seasonal evaporation is more important. However, fine tun-ng of irrigation scheduling, as affected by management practicesncluding mulching, may benefit from more accurate simulationn shorter time scales. The SOILWAT2 module also captured theffect of mulch on Es and soil water content in the upper soilayers.
Simulated cumulative Es during the whole crop season waseduced by mulch by 22 and 37 mm during the two years, closeo the values observed in the field. Simulated transpiration was 18nd 20 mm higher with mulch than without mulch in 2006–2007nd 2007–2008, respectively. This is consistent with our observa-ions (Balwinder-Singh et al., unpublished data), and Jones (1992)lso suggested that higher soil water content under mulch providesore available water for wheat growth, implying that the stomata
f wheat plants could open more freely. The simulated Es and tran-piration data suggest that the water saved from reduced Es underulch was not fully diverted to transpiration. We hypothesise that
nder mulch, water saved from reduced Es in the pre-anthesiseriod was used as transpiration in the post anthesis period, asvident from higher post anthesis biomass production with mulchFig. 2). This resulted in higher grain yield, as 70–90% of post anthe-
ter sowing
tion under mulch and non-mulch I2 in 2006–2007 (a and b) and 2007–2008 (c and
sis biomass production is deposited in the grain (Bidinger et al.,1977).
Total crop water use (ET) was underestimated in all treatments,by 5 to 23%. In irrigation treatments I1 and I2, the underestima-tion was only by 5 to 13%, with highest values in the mulchedtreatments, and was mainly due to lower simulated than observedtranspiration. This may be due to over-prediction of daily Es. Theunderestimation of transpiration by the model was greater withmulch than without mulch because in the field, the water savedfrom reduction in Es by mulch was fully diverted to transpiration. Inthe model, however, the saved water was stored in the soil profile,resulting in lower simulated transpiration with mulch. The greaterunderestimation of simulated ET in the water stressed treatmentsmay be because root growth slowed down or even ceased undersevere drought conditions, whereas field experiments have shownthat under deficit irrigation the plants attempted to exploit the soilprofile to satisfy evaporative demand by increasing the root depth(Gajri et al., 1993).
5. Conclusions
The APSIM model was calibrated and validated under cen-tral Punjab conditions for irrigated wheat grown in rotation withrice, with and without rice straw mulch, for a range of irrigationscheduling treatments. APSIM-Wheat performed well in terms ofsimulating biomass, yield, soil water dynamics, total ET and totalEs in contrasting rainfall years. However, the model did not cap-
ture the effect of mulch on crop development. The model alsounderpredicted yield under water deficit conditions, and did notpredict the effect of very low temperatures during the tilleringstage on suppression of biomass production. The results of the
1 d Crop
mnimceomwil
A
cJdfi
R
A
A
A
A
A
B
B
B
B
B
C
C
G
G
GG
H
H
H
J
2 Balwinder-Singh et al. / Fiel
odel evaluation indicate that the model can be used for prelimi-ary evaluation of crop management and cropping system options
nvolving wheat in north-west India. However, model improve-ents that capture the effect of mulch on soil temperature and
onsequent rates of crop growth and development are needed,specially for evaluation of intensive crop sequence options toptimise cropping system land and water productivity. Improve-ents in the simulation of daily Es, and of yield and biomass underater deficit, are also needed to improve the ability to identify
rrigation management to minimise irrigation input without yieldoss.
cknowledgements
We are grateful to the Australian Centre for International Agri-ultural Research (ACIAR) for support of the senior author through aohn Allwright Fellowship. We also thank Sarabjit Singh and Baljin-er Singh for their excellent technical assistance in conducting theeld experiments.
eferences
dams, J.E., Arkin, G.F., Ritchie, J.T., 1976. Influence of row spacing and straw mulchon 1st stage drying. Soil Sci. Soc. Am. J. 40, 436–442.
rora, V.K., Singh, H., Balwinder-Singh, 2007. Analyzing wheat productivityresponses to climatic, irrigation and fertilizer-nitrogen regimes in a semi-aridsub-tropical environment using the CERES-Wheat model. Agric. Water Manage.94, 22–30.
sseng, A., Keating, B.A., Fillery, I.R.P., Gregory, P.J., Bowden, J.W., Turner, N.C., Palta,J.A., Abrecht, D.G., 1998. Performance of the APSIM-Wheat model in WesternAustralia. Field Crops Res. 57, 163–179.
sseng, A., VanKeulen, H., Stol, W., 2000. Performance and application of the APSIMNwheat model in the Netherlands. Eur. J. Agron. 12, 37–54.
ydin, M., Yang, S.L., Kurt, N., Yano, T., 2005. Test of a simple model for estimat-ing evaporation from bare soils in different environments. Ecol. Modell. 182,91–105.
alwinder-Singh, Humphreys, E., Eberbach, P.E., Katupitiya, A., 2010. The par-titioning of evapotranspiration of irrigated wheat as affected by rice strawmulch. In: Soil Solutions for a Changing World. Proceedings of the 19thWorld Soil Science Congress, 1–6 August 2010, Brisbane, Australia (availableat http://www.iuss.org/19th%20WCSS/symposium/pdf/0628.pdf).
alwinder-Singh, Humphreys, E., Eberbach, P.L., Katupitiya, A., Singh, Y., Kukal, S.S.,2011. Growth, yield and water productivity of zero till wheat as affected by ricestraw mulch and irrigation schedule. Field Crops Res. 121, 209–225.
andyopadhyay, A., Bhadra, A., Raghuwanshi, N.S., Singh, R., 2008. Estimation ofmonthly solar radiation from measured air temperature extremes. Agric. For.Meteor. 148, 1707–1718.
idinger, F., Musgrave, R.B., Fischer, R.A., 1977. Contribution of stored pre-anthesisassimilate to grain-yield in wheat and barley. Nature 270, 431–433.
oatwright, G.O., Ferguson, H., Sims, J.R., 1976. Soil temperature around crown nodeinfluences early growth, nutrient-uptake, and nutrient translocation of springwheat. Agron. J. 68, 227–231.
hakraborty, D., Nagarajan, S., Aggarwal, P., Gupta, V.K., Tomar, R.K., Garg, R.N.,Sahoo, R.N., Sarkar, A., Chopra, U.K., Sarma, K.S.S., Kalra, N., 2008. Effect ofmulching on soil and plant water status, and the growth and yield of wheat(Triticum aestivum L.) in a semi-arid environment. Agric. Water Manage. 95,1323–1334.
haudhary, T.N., Chopra, U.K., 1983. Effect of soil covers on growth and yield ofirrigated wheat planted at two dated. Field Crops Res. 6, 293–304.
ajri, P.R., Prihar, S.S., Arora, V.K., 1993. Interdependence of nitrogen and irrigationeffects on growth and input-use efficiencies in wheat. Field Crops Res. 31, 71–86.
lanville, S.G., Freebairn, D.M., Silburn, D.M., 1984. Using curve numbers from simu-lated rainfall to describe the runoff characteristics of contour bay catchments. In:Conference on Agricultural Engineering, Bundaberg, Queensland, 27–30 August194, Institution of Engineers, Australia.
OP, 2006. Agriculture at a glance. Dept. of Agriculture, GOP, Chandigarh, India.uedira, M., Paulsen, G.M., 1988. Mediation of high temperature injury by roots and
shoots during reproductive growth of wheat. Plant Cell Environ. 11, 517–523.ay, R.K.M., Tunnicliffe, W.G., 1982. Leaf appearence and extension in field grown
winter wheat plants: the importance of soil temperature during vegetativegrowth. J. Agric. Sci. 99, 403–410.
umphreys, E., Kukal, S.S., Christen, E.W., Hira, G.S., Balwinder-Singh, Yadav, S.,Sharma, R.K., 2010. Halting the groundwater decline in north-west India—whichcrop technologies will be winners? Adv. Agron. 109, 155–217.
undal, S.S., Kaur, P., 1997. Application of the CERES-Wheat model to yield predic-tions in the irrigated plains of the Indian Punjab. J. Agric. Sci. 129, 13–18.
alota, S.K., Arora, V.K., 2002. Model-based assessment of water balance componentsunder different cropping systems in north-west India. Agric. Water Manage. 57,75–87.
s Research 124 (2011) 1–13
Jalota, S.K., Sood, A., Chahal, G.B.S., Choudhury, B.U., 2006. Crop water productivityof cotton (Gossypium hirsutum L.)-wheat (Triticum aestivum L.) system as influ-enced by deficit irrigation, soil texture and precipitation. Agric. Water Manage.84, 137–146.
Jones, H.G., 1992. Plant and Microclimate: A Quantitative Approach to Environ-mental Plant Physiology, second ed. The Press Syndicate of the University ofCambridge, New York.
Jones, C.A., Kiniry, J.R., 1986. CERES-Maize: A Simulation Model of Maize Growthand Development. Texas A&M University Press, College Station, TX, USA.
Kaspar, T.C., Bland, W.L., 1992. Soil-temperature and root-growth. Soil Sci. 154,290–299.
Keating, B.A., McCown, R.L., Creswell, H.P., 1995. Paddock-scale models andcatchment-scale problems: the role for APSIM in the Liverpool plains. In:Binning, P., Bridgman, H., Willimas, B. (Eds.), Proceedings of MODSIM ’95, Inter-national Congress on Modelling and Simulation. The University of Newcastle,NSW, Australia, pp. 158–165, 27–30 November.
Keating, B.A., Meinke, H., Probert, M.E., Huth, N.I., Hillis, I.G., 2001. Nwheat:documentation and performance of a wheat module for APSIM. In: TropicalAgronomy Memorandum. CSIRO Division of Tropical Agriculture, 306 CarmodyRd., St. Lucia, Qld 4067, Australia.
Keating, B.A., Carberry, P.S., Hammer, G.L., Probert, M.E., Robertson, M.J., Holzworth,D., Huth, N.I., Hargreaves, J.N.G., Meinke, H., Hochman, Z., McLean, G., Verberg,K., Snow, V., Dimes, J.P., Silburn, M., Wang, E., Brown, S., Bristow, K.L., Asseng,S., Chapman, S., Cown, Mc., Freebairn, R.L., Smith, D.M.C.J., 2003. An overviewof APSIM, a model designed for farming systems simulation. Eur. J. Agron. 18,267–288.
Keeney, D.R., Nelson, D.W., 1982. Nitrogen-inorganic forms. In: page, A.L., Miller,R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis—Part 2 Chemical and Micro-biological Properties. American Society of Agronomy, Soil Science Society ofAmerica, Madison, WI, USA, pp. 643–698.
Kuroyanagi, T., Paulsen, G.M., 1988. Mediation of high temperature injury byroots and shoots during reproductive growth of wheat. Plant Cell Environ. 11,517–523.
Ladha, J.K., Pathak, H., Gupta, R.K., 2007. Sustainability of the rice–wheat crop-ping system: issues, constraints, and remedial options. J. Crop Improv. 19,125–136.
Littleboy, M., Silburn, D.M., Freebairn, D.M., Woodruff, D.R., Hammer, G.L., Leslie, J.K.,1992. Impact of soil-erosion on production in cropping systems.1. Developmentand validation of a simulation-model. Aust. J. Soil Res. 30, 757–774.
Meinke, H., Rabbinge, R., Hammer, G.L., van Keulen, H., Jamieson, P.D., 1998.Improving wheat simulation capabilities in Australia from a cropping systemsperspective - II, Testing simulation capabilities of wheat growth. Eur. J. Agron.8, 83–99.
Miyasaka, S.C., Grunes, D.L., 1990. Root temperature and calcium level effects onwinter wheat forage:I shoot and root growth. Agron. J. 82, 236–242.
Priestley, C.H.B., Taylor, R.J., 1972. Assessment of surface heat-flux and evaporationusing large-scale parameters. Month. Weath. Rev. 100, 81–82.
Prihar, S.S., Jalota, S.K., Steiner, J.L., 1996. Residue management for reducingevaporation in relation to soil type and evaporativity. Soil Use Manage. 12,150–157.
Prihar, S.S., Gajri, P.R., Narang, R.S., 1974. Scheduling irrigations to wheat, using panevaporation. Indian J. Agric. Sci. 44, 567–571.
Probert, M.E., Keating, B.A., Thompson, J.P., Parton, W.J., 1995. Modelling water,nitrogen, and crop yield for a long-term fallow management experiment. Aust.J. Exp. Agric. 35, 941–950.
Probert, M.E., Dimes, J.P., Keating, B.A., Dalal, R.C., Strong, W.M., 1998. APSIM’s waterand nitrogen modules and simulation of the dynamics of water and nitrogen infallow systems. Agric. Syst. 56, 1–28.
Ritchie, J.T., 1972. Model for predicting evaporation from a row crop with incompletecover. Water Resour. Res. 8, 1204–1213.
Ritchie, J.T., Godwin, D.C., Otter, S., 1985. CERES-Wheat: a user-oriented wheat yieldmodel. Preliminary Documentation: AGRISTARS Publication No. YM-U3-04442-JSC-18892. Michigan State University, MI, USA.
Sharma, R.K., Chhokar, R.S., Jat, M.L., Samar, S., Mishra, B., Gupta, R.K., 2008.Direct drilling of wheat into rice residues: experiences in Haryana andWestern Uttar Pradesh. In: Humphreys, E., Roth, C.H. (Eds.), PermanentBeds and Rice-Residue Management for Rice–Wheat Systems in the Indo-Gangetic Plain. Proceedings of a Workshop. ACIAR Proceedings No. 127.PAU, Ludhiana, India from 7–9 September, 2006, pp. 147–158 (available athttp://www.aciar.gov.au/publication/term/18).
Sidhu, H.S., Singh, M., Blackwell, J., Humphreys, E., Bector, V., Singh, Y., Singh,M., Singh, S., 2008. Development of the Happy Seeder for direct drillinginto combine harvested rice. In: Humphreys, E., Roth, C.H. (Eds.), Per-manent Beds and Rice-Residue Management for Rice–Wheat Systems inthe Indo-Gangetic Plain. Proceedings of a Workshop. ACIAR ProceedingsNo. 127. PAU, Ludhiana, India, 7–9 September, pp. 159–170 (available athttp://www.aciar.gov.au/publication/term/18).
Sidhu, H.S., Manpreet, S., Humphreys, E., Yadvinder, S., Balwinder, S., Dhillon,S.S., Blackwell, J., Bector, V., Malkeet, S., Sarbjeet, S., 2007. The Happy Seederenables direct drilling of wheat into rice stubble. Aust. J. Exp. Agric. 47,844–854.
Sys, Ir.C., van Ranst, E., Debaveye, J., 1991. Land Evaluation. Part I. Principles in LandEvaluation and Crop Production Calculations. International Training Centre forPost-graduate Soil Scientists, University Ghent, Belgium.
Thorburn, P.J., Probert, M.E., Robertson, F.A., 2001. Modelling decomposition of sugarcane surface residues with APSIM-Residue. Field Crops Res. 70, 223–232.
d Crop
T
T
T
U
W
W
Balwinder-Singh et al. / Fiel
imsina, J., Humphreys, E., 2006. Performance of CERES-Rice and CERES-Wheat mod-els in rice & wheat systems: a review. Agric. Syst. 90, 5–31.
imsina, J., Godwin, D., Humphreys, E., Yadvinder, S., Bijay, S., Kukal, S.S., Smith,D., 2008. Evaluation of options for increasing yield and water productivity ofwheat in Punjab, ndia using the DSSAT-CSM-CERES-Wheat model. Agric. WaterManage. 95, 1099–1110.
ripathi, R.P., Singh, A.K., Mishra, R.K., 1985. Effect of mulch-induced soil thermalregime on wheat emergence. Field Crops Res. 10, 305–313.
domprasert, N., Li, P.H., Davis, D.W.A.H.M., 1995. Effects of roots temperatures onleaf gar exchange and growth at high air temperature in Phaseolus aculifoliusand Phaseolus vulgaris. Crop Sci. 35, 490–495.
allace, J.S., Lloyd, C.R., Sivakumar, M.V.K., 1993. Measurements of soil,plant and total evaporation from millet in Niger. Agric. For. Meteor. 63,
149–169.ang, E., Robertson, M.J., Hammer, G.L., Carberry, P.S., Holzworth, D., Meinke, H.,Chapman, S.C., Hargreaves, J.N.G., Huth, N.I., McLean, G., 2003. Developmentof a generic crop model template in the cropping system model APSIM. Eur. J.Agron. 18, 121–140.
s Research 124 (2011) 1–13 13
Singh, Y., Balwinder-Singh, Timsina, J., 2005. Crop residue management for nutrientcycling and improving soil productivity in rice-based cropping systems in thetropics. Adv. Agron. 85, 269–407.
Singh, Y., Sidhu, H.S., Singh, M., Humphreys, E., Kukal, S.S., Brar, N.K., 2008. Strawmulch, irrigation water and fertiliser N management effects on yield, wateruse and N use efficiency of wheat sown after rice. In: Humphreys, E., Roth,C.H. (Eds.), Permanent Beds and rice-residue management for rice–wheat sys-tems in the Indo-Gangetic Plain. Proceedings of a Workshop. ACIAR ProceedingsNo. 127. PAU, Ludhiana, India, 7–9 September 2006, pp. 171–181 (available athttp://www.aciar.gov.au/publication/term/18).
Singh, Y., Humphreys, E., Kukal, S.S., Balwinder-Singh, Kaur, A., Thaman, S., Prashar,A., Yadav, S., Timsina, J., Dhillon, S.S., Kaur, N., Smith, D.J., Gajri, P.R., 2009. Cropperformance in permanent raised bed rice–wheat cropping system in Punjab
India. Field Crops Res. 110, 1–20.Yunusa, I.A.M., Bellotti, W.D., Moore, A.D., Probert, M.E., Baldock, J.A., Miyan, S.M.,2004. An exploratory evaluation of APSIM to simulate growth and yield pro-cesses for winter cereals in rotation systems in South Australia. Aust. J. Exp.Agric. 44, 787–800.
