Content uploaded by Narsimha Rao
Author content
All content in this area was uploaded by Narsimha Rao on May 25, 2019
Content may be subject to copyright.
Content uploaded by Srinivasrao Ch.
Author content
All content in this area was uploaded by Srinivasrao Ch. on Jun 19, 2016
Content may be subject to copyright.
Content uploaded by Pragyan Kumari
Author content
All content in this area was uploaded by Pragyan Kumari on Nov 25, 2017
Content may be subject to copyright.
1 23
Environmental Modeling &
Assessment
ISSN 1420-2026
Environ Model Assess
DOI 10.1007/s10666-015-9462-6
Predicting Irrigated and Rainfed Rice Yield
Under Projected Climate Change Scenarios
in the Eastern Region of India
A.V.M.Subba Rao, Arun K.Shanker,
V.U.M.Rao, V.Narsimha Rao,
A.K.Singh, Pragyan Kumari,
C.B.Singh, Praveen Kumar Verma, et
al.
1 23
Your article is protected by copyright and
all rights are held exclusively by Springer
International Publishing Switzerland. This e-
offprint is for personal use only and shall not
be self-archived in electronic repositories. If
you wish to self-archive your article, please
use the accepted manuscript version for
posting on your own website. You may
further deposit the accepted manuscript
version in any repository, provided it is only
made publicly available 12 months after
official publication or later and provided
acknowledgement is given to the original
source of publication and a link is inserted
to the published article on Springer's
website. The link must be accompanied by
the following text: "The final publication is
available at link.springer.com”.
Predicting Irrigated and Rainfed Rice Yield Under Projected
Climate Change Scenarios in the Eastern Region of India
A. V. M. Subba Rao
2
&Arun K. Shanker
1
&V. U. M. Rao
2
&V. Narsimha Rao
2
&
A. K. Singh
3
&Pragyan Kumari
4
&C. B. Singh
5
&Praveen Kumar Verma
6
&
P. Vijaya Kumar
2
&B. Bapuji Rao
2
&Rajkumar Dhakar
2
&M. A. Sarath Chandran
7
&
C. V. Naidu
8
&J. L. Chaudhary
6
&Ch. Srinivasa Rao
7
&B. Venkateshwarlu
9
Received: 7 February 2014 /Accepted: 6 April 2015
#Springer International Publishing Switzerland 2015
Abstract Numerous estimates for the coming decades project
changes in precipitation resulting in more frequent droughts
and floods, rise in atmospheric CO
2
and temperature, exten-
sive runoff leading to leaching of soil nutrients, and decrease
in freshwater availability. Among these changes, elevated
CO
2
can affect crop yields in many ways. It is imperative to
understand the consequences of elevated CO
2
on the produc-
tivity of important agricultural crop species in order to devise
adaptation and mitigation strategies to combat impending cli-
mate change. In this study, we have modeled rice phenology,
growth phase, and yield with the BDecision Support System
for Agrotechnology Transfer (DSSAT) CERES rice model^
and arrived at predicted values of yield under different CO
2
concentrations at four different locations in Eastern India out
of which three locations were irrigated and one location was
rainfed. The ECHAM climate scenario, Model for Interdisci-
plinary Research on Climate (MIROC)3.0 climate scenario,
and ensemblemodels showed different levels of yield increase
with a clear reduction in yield under rainfed rice as compared
to irrigated rice. A distinct regional and cultivar difference in
response of rice yield to elevated CO
2
was seen in this study.
Results obtained by simulation modeling at different climate
change scenarios support the hypothesis that rice plant re-
sponses to elevated CO
2
are through stimulation of photosyn-
thesis. Realization of higher yields is linked with source sink
dynamics and partitioning of assimilates wherein sink capac-
ity plays an important role under elevated CO
2
conditions.
Keywords Climate change .Simulation modeling .
Elevated CO
2
.Sink capacity .Photosynthesis
1 Introduction
Global climate change will have a decisive impact on crop
production, and the prediction of the extent of this has
emerged as a major research priority during the past decade.
Numerous estimates for the coming decades project changes
in precipitation resulting in more frequent droughts and
floods, rise in atmospheric CO
2
and temperature, extensive
runoff leading to leaching of soil nutrients, and decrease in
freshwater availability. World agriculture faces overwhelming
*Arun K. Shanker
arunshank@gmail.com
1
Division of Crop Sciences, Central Research Institute for Dryland
Agriculture (CRIDA), Santoshnagar, Saidabad PO, Hyderabad 500
059, India
2
All India Coordinated Research Project on Agrometeorology, Central
Research Institute for Dryland Agriculture (CRIDA), Santoshnagar,
Saidabad PO, Hyderabad 500 059, India
3
Department of Agrometeorology, N.D. University of Agriculture and
Technology, Kumarganj, Faizabad 224 229, Uttar Pradesh, India
4
Department of Agricultural Physics, Birsa Agricultural University,
Kanke, Ranchi 834 006, Jharkhand, India
5
Department of Agronomy, C.S. Azad University of Agriculture and
Technology, Nawabganj, Kanpur 208 002, Uttar Pradesh, India
6
Department of Agro Meteorology, College of Agriculture, Indira
Gandhi Krishi Vishwavidyalaya (IGKV), Krishak Nagar,
Raipur 492004, Chhattisgarh, India
7
Central Research Institute for Dryland Agriculture (CRIDA),
Santoshnagar, Saidabad PO, Hyderabad 500 059, India
8
Andhra University, Vishakapatnam, Andhra Pradesh, India
9
Vasantrao Naik Marathwada Krishi Vidyapeeth Parbhani (VNMKV),
Parbhani 431402, Maharashtra, India
Environ Model Assess
DOI 10.1007/s10666-015-9462-6
Author's personal copy
challenges to meet rising demands for food, energy, and other
agricultural products, and this coupled with increasing popu-
lation and dwindling natural resources makes it imperative for
us to understand the consequences of climate change on the
productivity of important agricultural crop species. Climate
trends since 1980 were large enough in many countries to
offset a significant proportion of the potential increases in
average crop yields due to technological advances, CO
2
fer-
tilization, and other factors [9,18]. The AR5 report of Stocker
et al. [28] shows an estimated warming of 0.85 °C since 1880
with the fastest rate of warming in the Arctic. By the end ofthe
twenty-first century, the global surface temperature increase is
likely to exceed 1.5 °C relative to the 1850 to 1900 period for
most scenarios (except representative concentration pathways
(RCPs) RCP 2.6) and is likely to exceed 2.0 °C for many
scenarios (RCP 6.0 and RCP 8.5). In addition, precipitation
will become more variable, and episodes of extreme weather
will become more frequent and intense and last longer.
Projected climate change scenarios will affect crop productiv-
ity and also food security. Kii et al. [14] analyzed food avail-
ability and risk of hunger under the combined scenarios of
food demands and agroproductivity with and without climate
change by 2100 for the B2 scenario and concluded that future
food demand can be satisfied globally under all assumed com-
bined scenarios and also that a reduction of food access dis-
parity and increased progress in productivity are as important
as climate change mitigation for reducing the risk of hunger.
Crop production system involving complete input manage-
ment with recommended package of practices including
scheduling irrigation, nutrient application, and crop protection
is cost intensive for experimental data generation especially
when done at different sowing dates and multiple locations.
Appropriately validated crop simulation models could be used
to test many such combinations in a brief time with limited
expense. Such simulations can adequately describe relative
trends in yields caused by environmental variation [21,27].
Mechanistic models predict crop productivity based on
quantitative functional relationships essentially fundamental
to the processes studied such as photosynthesis, water rela-
tions, source sink relationship, and their interaction with
changing environmental variables. This makes these models
suitable for projecting the impact of predicted climate scenar-
ios on crop productivity; in addition, these models can be used
at multiple levels which include experimental field, regional,
national, and global. Local-gridded approaches to crop model-
ing using climate change scenarios can also assist to identify
suitable adaptation and mitigation measures. Lenzen et al. [16]
developed a modeling approach for examining selected
drivers of ecosystem functioning and agricultural productivity
and carried out a number of scenario projections of country-
level consumption, production, land use, energy use, green-
house gas emissions, species diversity, and agricultural pro-
duction up to 2050. Recently, application of the crop yield
simulation systems approach to climate change adaptation
and mitigation initiatives has been gaining popularity. This
is due to the increasing availability of information on the
processes that are affected by changing climate, thus helping
us to devise possible adaptation measures [6,20,30,31].
The effects of CO
2
fertilization on crop production have also
been widely studied and have been reported to reduce some
of the potential damages caused by the climatic impacts of
greenhouse gases but by significantly less than that indicated
in earlier research [5].
Rice (Oryza sativa L.) is one of the major crops in the
world. It is grown in diverse climatic zones and is the staple
diet of about 2.7 billion people. In India, it is cultivated in
about 150 million hectares, producing 132,013,000 metric
tons, which covers about 26 % of the global rice production
[29]. Climate change impacts on the yields of rice in relation
to temperature, carbon dioxide, and rainfall change have been
studied by using the crop growth model InfoCrop in the up-
lands of the Hill Tracts of Chittagong [3]. Rice cultivars re-
spond differently to elevated CO
2
concentrations, and CO
2
×
cultivar interaction has also been studied under open-field
conditions and it was found to increase yield in rice [7]. In
view of the above facts, we devised a study with the objectives
of calibrating and validating rice phenological stages with
special reference to anthesis days and physiological maturity
in different cultivars at different locations in Eastern India
including locations of the Indo-Gangetic Plains (IGP) of India
and to model the effect of elevated CO
2
on rice yield under
different climate change scenarios.
2 Materials and Methods
2.1 Study Area
In the present study, four representative locations were select-
ed, viz., Faizabad (IGP), Kanpur (IGP), Raipur, and Ranchi in
Eastern India. These locations varied in their geographical
placement and differed in climate as well as soils (Table 1).
Seasonal rainfall in these locations varied from 697.7 to
1148.3 mm. The mean maximum temperatures varied be-
tween 28.4 and 38 °C and the mean minimum temperatures
ranged between 15.5 and 27.5 °C. Rice is grown under an
irrigated ecosystem in Faizabad, Kanpur, and Raipur, whereas
at Ranchi, it is grown under a rainfed ecosystem. In the irri-
gated ecosystem, the amount of irrigation differed from loca-
tion to location.
2.2 Experimental Data
Experimental crop data for rice was collected from the date of
sowing experiments (early, normal, and late sowings) carried
out at Faizabad, Kanpur, Raipur, and Ranchi locations which
A.V.M.S. Rao et al.
Author's personal copy
are part of All India Coordinated Research Project on
Agrometeorology (AICRPAM), Indian Council of Agricultur-
al Research (ICAR), Central Research Institute for Dryland
Agriculture (CRIDA), Hyderabad, India. A minimum data
set of the summer (kharif) rice crop which includes crop man-
agement data, viz., cultivar information, crop duration, date of
sowing, date of anthesis or 50 % flowering, date of physio-
logical maturity, application schedule, and amount of fertilizer
and irrigation, is given in Table 2. Soil profile data comprising
the physical and chemical properties of the soil up to 60 cm
depth with an interval of 5 cm was collected from the study
area of the four locations. Further, locationwise daily weather
data (solar radiation, maximum and minimum temperatures,
and rainfall) was collected and arranged in the CERES rice
model weather input format. All this crop management data
along with soil profile and daily weather data sets of the se-
lected locations was converted to model data sets for further
simulation process.
2.3 Decision Support System for Agrotechnology Transfer
CERES Rice Model
The CERES rice model in decision support system for
agrotechnology transfer (DSSAT) developed by Hoogenboom
et al. [8] is a process-based model that can simulate the growth
and development of cereal crops under varying weather, soil,
and management levels. It simulates soil water balance and
water use by the crop and soil nitrogen transformations and
uptake by the crop besides growth and different phenophases
of the crop. Crop growth rate (CGR) is simulated by
employing a carbon balance approach in a source sink system
[23] and crop duration through thermal time concept [11]. A
schematic representation of the model is given in Fig. 1.The
model simulates total biomass of the crop as the product of the
growth duration and average growth rate. The model uses
phenology, nitrogen and water input, and growth and devel-
opment as components wherein day length and temperature
are taken into consideration for the phenological stages. Bio-
mass production is taken into account as a result of
partitioning of assimilates; in addition, source sink relation-
ship is also considered in the model. The simulation of yields
at the process level involves the prediction of these two im-
portant processes. The yield of the crop is the fraction of total
biomass partitioned to grain.
2.4 Model Calibration and Validation
Calibration is adjustment of the system parameters so that
simulated results reach a predetermined level, usually that of
an observation. Genotype coefficients of the rice cultivars
‘Sarjoo-52 (Faizabad), NDR359 (Kanpur), Mahamaya (Rai-
pur), and Vandana (Ranchi)’were calibrated and validated
using the experimental data on phenology and grain yield
Tab le 1 Agroclimatic and edaphic details of the selected locations
Station Latitude Longitude Altitude (m) Cultivar Soil Ecotype Period of experiment T
max
(°C) T
min
(°C) Rainfall (mm) Sunshine (h)
Faizabad 26° 47′N82°12′E 113 Sarjoo-52 Silty loam Irrigated 2008–2012 33.8 19.4 812 7.2
Kanpur 26° 39′35″N80°18′25″E 125.9 NDR359 Sandy loam Irrigated 2008–2012 31.3 18.7 867 5.9
Raipur 21° 15′00″N81°41′00″E 289 Mahamaya Matasil soil Irrigated 2007–2011 32.7 19.7 1180 6.7
Ranchi 23° 17′00″N85°10′00″E 289 Vandana Loamy Rainfed 2007–2012 28.6 16.7 1397 6.02
Predicting Rice Yield Under Climate Change Scenarios in India
Author's personal copy
Tab le 2 Cultivation and input application details of varieties in different locations
Location Cultivar Dates of sowing Dates of irrigation application Amount of irrigation Dates of fertilizer application Type of fertilizer Amount of fertilizer (N, P, and K) application (kg ha
−1
)
Faizabad Sarjoo-52 Early 24th July 60 mm 5th July Urea, DAP, MOP 50 % N +P and K 120:60:60
19th August 29th July 25 % N
9thSeptember 2ndSeptember 25%N
Normal 10th August 15th July Urea, DAP, MOP 50 % N +P and K
25th August 10th August 25 % N
10th September 13th September 25 % N
Late 15th August 25th July Urea, DAP, MOP 50 % N +P and K
30th August 17th August 25 % N
15th September 15th September 25 % N
Kanpur NDR359 Early 12th July Check basin 05th July Urea, DAP, MOP 25 % N +P and K 211:132:102
11th August 3rd August 50 % N
26th August 04th September 25 % N
15th September
Normal 22nd July 15th July Urea, DAP, MOP 25 % N +P and K
21st August 13th August 50 % N
05th September 14th September 25 % N
25th September
Late 1st August 25th July Urea, DAP, MOP 25 % N +P and K
31st August 23rd August 50 % N
15th September 24th September 25 % N
5th October
Raipur Mahamaya Early 16th July 75 mm 15th July Urea, DAP, MOP 33 % (N+P and K) 80:60:40
13th October 12th August
14th September
Normal 26th July 25th July Urea, DAP, MOP 33 % (N +P and K)
23rd October 25th August
23th September
Late 03rd August 5th August Urea, DAP, MOP 34 % (N+P and K)
31st October 4th September
2nd October
Ranchi Vandana Early No irrigation No irrigation 20th June Urea, SSP, MOP 25 % N +P and K 131:188:35
7th July 50 % N
26th July 25 % N
Normal 30th June Urea, SSP, MOP 25 % N+ P and K
20th July 50 % N
5th August 25 % N
Late 10th July Urea, SSP, MOP 25 % N +P and K
30th July 50 % N
15th August 25 % N
A.V.M.S. Rao et al.
Author's personal copy
for different years (Table 3). Statistical methods were selected
to compare the results from simulation and observation. Mod-
el performance evaluation is presented by the root mean
square error (RMSE) and D-index [32,33].
RMSE ¼X
n
i¼1
SiObi
ðÞ
2
n
()
0:5
D¼1−Xn
i¼1Si−Obi
ðÞ
2
Xn
i¼1SiObavg þ
jj
SiObavg
2
where S
i
and Ob
i
are the model simulated and experimental
observation points, respectively. Ob
avg
is the average of ex-
perimental observations and nis the number of observations.
2.5 Climate Change Scenarios
Climate change influences rice crop mainly through increased
atmospheric CO
2
, temperature, and change in rainfall. The
Climate Change Agriculture and Food Security (CCAFS) In-
stitute under the CGIAR system has hosted a web site for
providing the downscaled projection data on point basis
(http://gismap.ciat.cgiar.org/MarkSimGCM/), through Mark
Sim™DSSAT weather file generator [10]. This generator
converts the downscaled weather data from global climate
models (GCMs) to DSSAT weather input file format.
Projected data collected from two GCMs, viz., ECHAM5
[24] and Model for Interdisciplinary Research on Climate
(MIROC)3.2 (K-1 Model Developers, 2004) data for 2020,
2040, and 2080, were used for simulation.
2.5.1 The ECHAM5 Climate Model
This climate model has been developed from the European
Center for Medium Range Weather Forecasting (ECMWF)
forecast atmospheric model and has a comprehensive param-
eterization package which allows the model to be used for
climate simulations. The model is a spectral transform model
with 19 atmospheric layers and derived from experiments per-
formed with spatial resolution T42 (which approximates to
about 2.8° longitude/latitude resolution). ECHAM5 is the cur-
rent generation in the line of ECHAM models [25]. A sum-
mary of developments regarding model physics in ECHAM
and a description of the simulated climate obtained with the
uncoupled ECHAM model are given in Roeckner et al. [26].
2.5.2 MIROC3.2
The MIROC, updated in 2004 to version 3.2, was jointly
developed in Japan by the Centre for Climate System
Fig. 1 Schematic representation
of the rice growth simulation
model CERES rice
Tabl e 3 Genetic coefficients of
different rice cultivars grown at
different locations
Station Cultivar/genotype P1 P2R P5 P2O G1 G2 G3 G4
Faizabad Sarjoo-52 680 163 365 12.3 70 0.0275 1 1
Kanpur NDR359 395.9 211.9 352.3 10.8 60 0.028 0.7 1.05
Raipur Mahamaya 618.4 233.9 384 11.9 56.6 0.026 1 1
Ranchi Vandana 258 60.2 216.5 12.5 72.4 0.028 1 1
Predicting Rice Yield Under Climate Change Scenarios in India
Author's personal copy
Research (CCSR), University of Tokyo; the National Institute
for Environmental Studies (NIES); the Frontier Research Cen-
tre for Global Change (FRCGC); and the Japan Agency for
Marine-Earth Science and Technology (JAMSTEC) [13].
This coupled model is comprised of the following models:
Atmospheric model: AGCM5.7b
Oceanic and sea ice model: COCO3.3
Land surface model: minimal advanced treatments of sur-
face interaction and runoff (MATSIRO)
Coupling model: MIROC3.2
2.5.3 Ensemble Model
The ensemble mean of four models, viz., CNRM-CM3,
CSIRO-Mk3_5, along with the above-described ECHAM5
and MIROC3.2, was used for the same years. CNRM-CM3
is a global coupled system, the third version of the ocean-
atmosphere model initially developed at CERFACS (Tou-
louse, France), and then regularly updated at the Center Na-
tional Weather Research. CNRM-CM3 also now includes a
parameterization of the homogeneous and heterogeneous
chemistry of ozone and a sea ice model, and it corresponds
to a resolution of about 2° in longitude, the latitudinal resolu-
tion varying from 0.5° near the equator to roughly 2° in polar
regions. CSIRO-Mk3_5 is a climate system model containing
a comprehensive representation of the four major components
of the climate system (atmosphere, land surface, oceans, and
sea ice). There are simulations for a range of scenarios avail-
able for this model. This simulation uses scenario SRESA2
which represents the SRESA2 scenario 2001 to 2100. The
scenario includes standard daily and monthly meteorological
and monthly oceanographic variables. It is also a contribution
to the WCRP CMIP3 multimodel database and meets their
formatting standards [4].
3Results
3.1 Rice Growth and Yield Calibration and Validation
Rice calibration in relation to growth and development in
terms of anthesis and physiological maturity at four locations,
viz., Ranchi, Faizabad, Kanpur, and Raipur, is shown in
Fig. 2.Figure2a shows the relationship between simulated
and observed anthesis days from 50 to 110 days and Fig. 2b
shows physiological maturity between 80 and 140 days. The
simulated and observed anthesis days and physiological ma-
turity days, respectively, in the case of Ranchi did not overlap
with the other locations, namely Faizabad, Kanpur, and Rai-
pur. On the other hand, rice calibration for anthesis and matu-
rity in the other locations, viz., Faizabad, Kanpur, and Raipur,
exhibited overlap with anthesis clustering between 85 and
105 days, with the earliest anthesis among these locations
being in Faizabad followed by Kanpur and Raipur. Rice cal-
ibrated in Kanpur exhibited a wider range in the number of
days for anthesis, starting from 90 to 105 days. A similar trend
was seen in the case of physiological maturity in the calibra-
tion of rice for the three locations, i.e., Faizabad, Kanpur, and
Raipur. Among these three locations, Faizabad rice came ear-
liest to maturity at 100 days and extended up to 125 days,
whereas two samples of Kanpur and Raipur rice were clus-
tered around 130 days showing late physiological maturity.
Maturity in all the locations did not extend beyond 132 days.
Calibration of rice in relation to growth and development was
closest to a linear trend in the case of locations Kanpur and
Faizabad and was slightly away in Raipur. A similar trend was
seen in the case of physiological maturity as well, although the
distance from the linear was a little more than that observed in
the case of anthesis. Ranchi also exhibited similar trends in the
case of both days to anthesis and physiological maturity. Rice
calibration in relation to simulated and observed yield (kg
ha
−1
)isshowninFig.2c. Rice yield in Ranchi was signifi-
cantly less compared to the other three locations, Faizabad,
Kanpur, and Raipur, ranging from 2200 to about
2800 kg ha
−1
. The yields in the other locations were more than
1000 kg ha
−1
and above the yield at Ranchi. The highest yield
was seen in Kanpur and the lowest yield among the three
locations other than Ranchi was seen in Faizabad. Yield in
Raipur was clustered between 4200 and 4800 kg ha
−1
.On
the other hand, yield in Kanpur was spread apart between
4800 and 6400 kg ha
−1
. Raipur exhibited closest to linearity
among the locations. A trendsimilar to the one observed in the
case of anthesis and maturity was also observed in yield in the
case of Ranchi.
Validation of data in rice in the four locations—Ranchi,
Faizabad, Kanpur, and Raipur—is depicted in Fig. 3. Valida-
tion of observed and simulated data on days to anthesis in
Ranchi followed a linear pattern and the anthesis was ob-
served between 65 and 70 days (Fig. 3a). Data were clustered
around these days and exhibited linearity, with only three
sample points being outliers. The other three locations were
closely clustered between 85 and 105 days. The validation
data for observed anthesis days in rice showed maximum var-
iation and spread in Kanpur with anthesis extending from 87
to about 103 days. On the other hand, the locations, namely
Faizabad and Raipur, did not exhibit a high degree of spread in
anthesis days, and they varied from 85 to 85 and 93 to
103 days in Faizabad and Raipur, respectively.
Similar to the validation data in rice for anthesis days in the
four selected locations, the validation data on physiological
maturity also showed that rice at Ranchi came to maturity
significantly earlier than the other three locations by about
10–15 days. In Raipur, unlike anthesis days (Fig. 3b), physi-
ological maturity was more spread out, ranging from 119 to
A.V.M.S. Rao et al.
Author's personal copy
130 days, whereas in the case of Faizabad, the same type of
clustering was seen in anthesis days. The highest degree of
variation in the days to maturity was seen in Kanpur starting
from as early as 118 days to as late as 131 days.
The relationship between simulated and observed rice yields
in the validation data in the four locations, viz., Ranchi, Faiza-
bad, Kanpur, and Raipur, is shown in Fig. 3c. Unlike that seen in
the anthesis and maturity, the difference between the outlier
Ranchi and the other states was not evident. The three loca-
tions—Faizabad, Kanpur, and Raipur—did not exhibit close
clustering as seen in anthesis and maturity. In this case, they
exhibited a high degree of spread. Ranchi showed the minimum
yield with a range of 1200 to near 3000 kg ha
−1
. The highest
yield observed in Ranchi was less than the lowest observed in
the other three locations, namely Faizabad, Kanpur, and Raipur.
Similar to the observations in anthesis days, the location Kanpur
exhibited a wide range in yield ranging from 4000 to
8000 kg ha
−1
. On the other hand, Faizabad and Raipur did not
exhibit this degree of variation as observed in Kanpur. Data
from Faizabad was clustered closely and was deviating from
linearity, whereas data from Raipur was spread out but close
to linearity. Table 4shows the rice anthesis day calibration at
Ranchi, Faizabad, Kanpur, and Raipur, and a 2-day difference
between observed and simulated was seen in the cultivar
Vandana in Ranchi in anthesis days. The cultivar Vandana came
to anthesis at the earliest as compared with the other three vari-
eties. A maximum difference was seen in the variety Mahamaya
which was 10 days with an RMSE value of 10.739. With regard
to rice physiological maturity day calibration (Table 5), again, it
was seen that cultivar Vandana attained physiological maturity
as compared to the other varieties in the other locations at
93 days and varied only 3 days from the simulated value
Fig. 2 Rice calibration in relation
to growth and development in
terms of anthesis (a),
physiological maturity (b), and
yield (c) at four locations, viz.,
Ranchi (Vandana), Faizabad
(Sarjoo-52), Kanpur (NDR359),
and Raipur (Mahamaya)
Predicting Rice Yield Under Climate Change Scenarios in India
Author's personal copy
because the variety Vandana is a short duration variety with
100-day duration. The physiological maturity date of this culti-
var at Ranchi was earlier than the anthesis days of other varieties
in the other three locations. In Kanpur and Raipur, varieties
NDR359 and Mahamaya, respectively, came to observed phys-
iological maturity 1 day apart at 126 and 127 days, and the
simulated values for theses varieties were 119 and 117 days,
respectively. Table 6shows rice yield calibration at Ranchi,
Fig. 3 Rice validation in relation
to growth and development in
terms of anthesis (a),
physiological maturity (b), and
yield (c) at four locations, viz.,
Ranchi (Vandana), Faizabad
(Sarjoo-52), Kanpur (NDR359),
and Raipur (Mahamaya)
Tabl e 4 Rice anthesis day calibration at Ranchi, Faizabad, Kanpur, and
Raipur
Station Cultivar Anthesis days RMSE D-stat
Observed Simulated
Faizabad Sarjoo-52 91 88 4.435 0.537
Kanpur NDR359 97 92 4.69 0.864
Raipur Mahamaya 98 88 10.739 0.322
Ranchi Vandana 66 68 5.627 0.202
Table 5 Rice maturity day calibration at Ranchi, Faizabad, Kanpur,
and Raipur
Station Cultivar Maturity days RMSE D-stat
Observed Simulated
Faizabad Sarjoo-52 116 113 6.272 0.382
Kanpur NDR359 126 119 7.767 0.628
Raipur Mahamaya 127 117 10.296 0.213
Ranchi Vandana 93 90 5.745 0.398
A.V.M.S. Rao et al.
Author's personal copy
Faizabad, Kanpur, and Raipur, and the Ranchi cultivar Vandana
recorded the least yield and the difference in yield between
observed and simulated was only 11 kg ha
−1
. The highest ob-
served yield of 5518 kg ha
−1
was recorded by variety NDR359
in Kanpur. The difference between observed and simulated
yield in the case of locations Faizabad, Kanpur, and Raipur
was much higher as compared to Ranchi.
4Discussion
Calibration was done with the independent data sets of four rice
cultivars, viz., Ranchi (Vandana), Faizabad (Sarjoo-52), Kanpur
(NDR359), and Raipur (Mahamaya), for different genetic coef-
ficients with the aim to characterize the rice performance. Ac-
curacy in simulation of yield, phenology, and growth requires
the use of accurate genetic coefficients [19,22]. These coeffi-
cients were adjusted and recalibrated until there was a high
degree of similarity between the observed and simulated dates
of anthesis, physiological maturity, and grain yield. Model val-
idation was done with observations on anthesis days, physio-
logical maturity days, and grain yield. Calibration and valida-
tion showed that there was good agreement between predicted
and observed values for all the phenology and yield parameters.
It was seen that in rice calibration at Ranchi, the anthesis days
were between 60 and 75 days and physiological maturity was
between 85 and 100 days. Ranchi rice was grown under upland
conditions and was not irrigated or flooded; hence, it was seen
that the growth and development of rice in the location Ranchi
was markedly different from the other locations wherein phe-
nology was delayed with respect to Ranchi. This could be be-
cause there is hastening of phenology under nonirrigated con-
ditions in rice possibly to adjust source sink partitioning so that
maximum source strength and sink capacity can be achieved
under given conditions. This can also be explained due to the
fact that, in general, time of day of flowering is regulated at the
level of the individual spikelet because anthesis events on a
panicle are spread over an extended period, probably due to
different physiological ages and topological positions among
spikelets [12]. Changes in environmental conditions preceding
the anthesis event and immediately after the anthesis event can
influence the time of flowering. This suggests the involvement
of regulatory processes occurring prior to anthesis as well as
after the anthesis event. It was seen in this study that physiolog-
ical maturity varied significantly in the cultivar that came to
early maturity in Ranchi. It has been observed in this study that
Ranchi cultivars advanced its phenology and growth phase un-
der growing conditions so that its flowering time effectively
escaped severe water stress which is frequently seen after the
end of the rainy season. This ability to maintain growth during
drought can contribute to yield determining processes.
4.1 Genetic Coefficients
Genetic coefficients of the four baseline cultivars used in the
study with respect to rice grown in different locations are
given in Table 3. The data shows P1—the time period
(expressed as growing degree days [GDD] in °C above a base
temperature of 9 °C) from seedling emergence during which
the rice plant is not responsive to changes in photoperiod. This
period is also referred to as the basic vegetative phase of the
plant. P2O is the critical photoperiod or the longest day length
(in hours) at which the development occurs at a maximum
rate. At values higher than P2O, developmental rate is slowed;
hence, there is a delay due to longer day lengths. P2R is the
extent to which phasic development leading to panicle initia-
tion is delayed (expressed as GDD in °C) for each hour in-
crease in photoperiod above P2O. P5 is the time period in
GDD (°C) from beginning of grain filling (3 to 4 days after
flowering) to physiological maturity with a base temperature
of 9 °C. G1 is the potential spikelet number coefficient as
estimated from the number of spikelets per gram of main culm
dry weight (less lead blades and sheaths plus spikes) at anthe-
sis. G2 is the single grain weight (g) under ideal growing
conditions, i.e., nonlimiting light, water, and nutrients and
absence of pests and diseases. G3 is the tillering coefficient
(scaler value) relative to IR64 cultivar under ideal conditions.
A higher tillering cultivar would have coefficient greater than
1. G4 is the temperature tolerance coefficient.
4.2 Rice Yield Prediction in Different Scenarios
Rice yield variation as reduction or increase from baseline
in the MIROC3.0 climate scenario under different CO
2
levels in 2020, 2040, and 2080 at different dates of sow-
ing—D1, D2, and D3—wherein D1 is the early-sown, D2
is the normal-sown, and D3 is the late-sown rice, is shown
in Fig. 4.ElevatedCO
2
concentration of 680 ppm showed
the highest yield in all the locations in all the dates of
sowing studied and in all the 3 years 2020, 2040, and
2080, and a reduction in yield was seen to be least in
Ranchi wherein the yields expressed as percentage differ-
ence over baseline were seen to be negative. Normal and
late date of sowing at an elevated CO
2
concentration of
420 ppm showed high reduction in yield in 2080 in
Tabl e 6 Rice yield calibration at Ranchi, Faizabad, Kanpur, and Raipur
Station Cultivar Yield RMSE D-stat
Observed Simulated
Faizabad Sarjoo-52 3822 4208 490.626 0.546
Kanpur NDR359 5518 5979 686.676 0.543
Raipur Mahamaya 4445 4508 83.694 0.909
Ranchi Vandana 2359 2370 553.508 0.364
Predicting Rice Yield Under Climate Change Scenarios in India
Author's personal copy
Ranchi, whereas this was not the case in the other years
with all the dates of sowing. In general, in all the locations,
all concentrations of CO
2
in all dates of sowing at 420 ppm
showed the least increase over baseline with reduction seen
in Ranchi. In Faizabad, there was a high increase in yield
from 30 to 45 % in all the concentrations of elevated CO
2
.
Fig. 4 Rice yield variation over
baseline (%) in MIROC3.0
climate scenario model under
different CO
2
levels (420, 550,
and 680 ppm) in 2020, 2040, and
2080 at different dates of
sowing—D1 (early sown), D2
(normal sowing), and D3 (late
sowing)—at Raipur (Mahamaya)
(a), Faizabad (Sarjoo-52) (b),
Kanpur (NDR359) (c), and
Ranchi (Vandana) (d)
A.V.M.S. Rao et al.
Author's personal copy
Unlike Faizabad and Kanpur, Raipur showed a decreasing
trend in yield as the date of sowing advanced.
Rice yield variation as reduction or increase from base-
line in ECHAM5 climate scenario under different CO
2
levels in 2020, 2040, and 2080 at different dates of sow-
ing—D1, D2, and D3—is shown in the ECHAM5 scenario
model is shown in Fig. 5. Unlike the MIROC3.0 climate
scenario, this model shows a decline in yield in 2080 in
three locations, namely Raipur, Kanpur, and Ranchi. An
elevated concentration of CO
2
at 680 ppm showed the max-
imum increase in all dates of sowing in 2080, and in Faiz-
abad, it was seen as negative as it showed the minimum
reduction as compared to the other concentrations. In Faiz-
abad, which was the only location where there was an
Fig. 5 Rice yield variation over
baseline (%) in ECHAM climate
scenario model under different
CO
2
levels (420, 550, and
680 ppm) in 2020, 2040, and
2080 at different dates of
sowing—D1 (early sown), D2
(normal sowing), and D3 (late
sowing)—at Raipur (Mahamaya)
(a), Faizabad (Sarjoo-52) (b),
Kanpur (NDR359) (c), and
Ranchi (Vandana) (d)
Predicting Rice Yield Under Climate Change Scenarios in India
Author's personal copy
increase in predicted yield, it was seen that the late sowing
date (D3) showed a high increase in yield under all three
elevated CO
2
concentrations.
Rice yield variation as reduction or increase from baseline
in the climate scenario ensemble models is shown in Fig. 6.In
general, CO
2
concentration of 420 ppm in all the locations
Fig. 6 Rice yield variation over
baseline (%) in ensemble climate
scenario models under different
CO
2
levels (420, 550, and
680 ppm) in 2020, 2040, and
2080 at different dates of
sowing—D1 (early sown), D2
(normal sowing), and D3 (late
sowing)—at Raipur (Mahamaya)
(a), Faizabad (Sarjoo-52) (b),
Kanpur (NDR359) (c), and
Ranchi (Vandana) (d)
A.V.M.S. Rao et al.
Author's personal copy
showed a marked decline in predicted yield, with the maxi-
mum reduction seen in 2080 in Ranchi. Late-sown rice in
Faizabad and Kanpur in 2040 and 2080 showed a yield in-
crease up to 40 % in all the three elevated CO
2
concentrations,
with higher elevation showing a higher percentage increase. In
Ranchi, the ensemble model showed a decrease in yield in all
the dates of sowing and all the years at all the elevated con-
centrations of CO
2
except 680 ppm in 2020 wherein at all
sowing dates a slight increase over the baseline yield was
predicted. The increase in CO
2
concentration increased yield
in general in all the scenarios simulated, and this is agreement
with many other researchers who have reported increasing rice
yield with increasing CO
2
level [1,2,15,17,34]. Results
obtained by simulation and scenarios support the hypothesis
that rice plant responses to elevated CO
2
are through stimula-
tion of photosynthesis and eventual realization of higher
yields. A distinct regional and cultivar difference in response
of rice yield to elevated CO
2
was seen in this study, and this
can be explained by the fact that while crop yield normally
responds positively to increased atmospheric CO
2
concentra-
tion, in our study, it was seen that the response depended on
the location and cultivar and, to a large extent, on the input
given as well as on the environmental conditions. Our re-
sults are in confirmation with Hasegawa et al. [7] who re-
ported that rice cultivars respond differently to elevated CO
2
concentrations. Here we see that CO
2
×cultivar interaction
tested by model prediction showed a yield increase which
could be due to larger sink capacity under higher CO
2
con-
centrations. Therefore, we can infer that inspite of increasing
CO
2
as predicted by climate change scenarios, regional and
local yield response to elevated CO
2
will vary due to differ-
ences in the local climate and the input used.
5Conclusions
Rice is one of the major crops in India. The ongoing climat-
ic aberrations may impact the yields and, hence, the food
security of the country. The present study was conducted by
selecting four rice-growing locations in Eastern India includ-
ing locations of the IGP of India using the DSSAT CERES
rice model, and a simulation was carried out by incorporat-
ing the future climate change scenarios from individual and
ensemble models. It was found that the rice yields are im-
pacted at all locations, viz., Kanpur, Faizabad, Raipur, and
Ranchi. Ranchi was seen to have maximum losses in terms
of yield because of the rainfed rice growing ecosystem. In-
crease in CO
2
concentrations increased the yields of rice in
all the locations except Ranchi. In Ranchi, yields were not
influenced by any concentration of elevated CO
2
except at
680 ppm during the 2020 scenario. Hastening of phenology
under rainfed conditions in rice is a possible physiological
mechanism to regulate source sink partitioning so that
maximum source strength and sink capacity can be attained
under given conditions. Adjusting the date of sowings can
become helpful as an adaptation strategy to identify the op-
timum date of sowing under future climate change scenarios
and, in turn, increase productivity.
References
1. Baker, J. T., Allen, L. H., & Boote, K. J. (1992). Temperature
effects on rice at elevated CO
2
concentration. Journal of
Experimental Botany, 43,959–964.
2. Baker, J. T., Allen, L. H., Jr., & Boote, K. J. (1996). Assessment of
rice response to global climate change: CO
2
and temperature. In G.
W.Koch&H.A.Mooney(Eds.),Carbon dioxide and terrestrial
ecosystems (pp. 256–282). San Diego: Academic.
3. Bala, B. K., & Hossain, M. A. (2013). Modeling of ecological
footprint and climate change impacts on food security of the Hill
Tracts of Chittagong in Bangladesh. Environmental Modeling &
Assessment, 18,39–55.
4. Collier, M. A., Dix, M. R., & Hirst, A. C. (2007). CSIRO Mk3
climate system model and meeting the strict IPCC AR4 data re-
quirements. MODSIM07 International Congress on Modelling and
Simulation: land, water & environmental management: integrated
systems for sustainability: Christchurch, 10–13 December 2007:
Proceedings, Christchurch, N.Z.
5. Darwin, R., & Kennedy, D. (2000). Economic effects of CO2 fer-
tilization of crops: transforming changes in yield into changes in
supply. Environmental Modeling & Assessment, 5,157–168.
6. Falloon, P., & Betts, R. (2010). Climate impacts on European agri-
culture and water management in the context of adaptation and
mitigation—the importance of an integrated approach. Science of
the Total Environment, 408(23), 5667–5687.
7. Hasegawa, T., Sakai, H., Tokida, T., Nakamura, H., Zhu, C., Usui,
Y., & Makino, A. (2013).Rice cultivar responses to elevated CO
2
at
two free-air CO
2
enrichment (FACE) sites in Japan. Functional
Plant Biology, 40(2), 148–159.
8. Hoogenboom, G., Wilkens, P. W., Thornton, P. K., Jones, J. W.,
Hunt, L. A., & Imamura, D. T. (1999). Decision support system for
agrotechnology transfer 3.5. In G. Hoogenboom, P. W. Wilkens, &
G. Y. Tsuji (Eds.), DSSAT version 3, vol. 4 (ISBN 1-886684-04-9)
(pp. 1–36). Honolulu: University of Hawaii.
9. IPCC. (2007). Contribution of workinggroup I to the fourth assess-
ment report of the intergovernmental panel on climate change.
Cambridge: Cambridge University Press.
10. Jones P.G., Thornton P.K. and Heinke J. (2011). Generating char-
acteristic daily weather data using downscaled climate model data
from the IPCC fourth assessment. https://hc.box.net/shared/
f2gk053td8.
11. Jones, J. W., Hoogenboom, G., Porter, C. H., Boote, K. J.,
Batchelor, W. D., Hunt, L. A., Wilkens, P. W., Singh, U.,
Gijsman, A. J., & Ritchie, J. T. (2003). The DSSAT cropping sys-
tem model. European Journal of Agronomy, 18,235–265.
12. Julia, C., & Dingkuhn, M. (2012). Variation in time of day of an-
thesis in rice in different climatic environments. European Journal
of Agronomy, 43,166–174.
13. K-1 Model Developers, 2004. K-1 coupled model (MIROC) de-
scription. K-1 technical report 1, Hasumi, H., Emori, S. (Eds.),
Center for Climate System Research, University of Tokyo, Tokyo,
Japan
Predicting Rice Yield Under Climate Change Scenarios in India
Author's personal copy
14. Kii, M., Akimoto, K., & Hayashi, A. (2013). Risk of hunger under
climate change, social disparity, and agroproductivity scenarios.
Environmental Modeling & Assessment, 18,299–317.
15. Kim, H. Y., Lieffering, M., Kobayashi, K., Okada, M., & Miura, S.
(2003). Seasonal change in the effects of elevated CO
2
on rice at
three levels of nitrogen supply: a free air CO
2
enrichment (FACE)
experiment. Global Change Biology, 9,826–837.
16. Lenzen, M., Dey, C., Foran, B., Widmer-Cooper, A., Ohlemüller,
R., Williams, M., & Wiedmann, T. (2013). Modelling interactions
between economic activity, greenhouse gas emissions, biodiversity
and agricultural production. Environmental Modeling &
Assessment, 18,377–416.
17. Liu, H., Yang, L., Wang, Y., Huang, J., Zhu, J., Yunxia, W., Dong,
G., & Liu, G. (2008). Yield formation of CO
2
-enriched hybrid rice
cultivar Shanyou under fully open-airfield conditions. Field Crops
Research, 108,93–100.
18. Lobell, D. B., Schlenker, W., & Costa-Roberts, J. (2011). Climate
trends and global crop production since 1980. Science, 333,616–
620.
19. Mavromatis, T., Boote, K. J., Jones, J. W., Irmak, A., Shinde, D., &
Hoogenboom, G. (2001). Developing genetic coefficients for crop
simulation models with data from crop performance trials. Crop
Science, 41(1), 40–51.
20. Nelson, G. C., & Shively, G. E. (2014). Modeling climate change
and agriculture: an introduction to the special issue. Agricultural
Economics, 45(1), 1–2.
21. Penning de Vries, F., Jansen, D., ten Berge, H., & Bakema, A.
(1989). Simulation of eco-physiological processes of growth in sev-
eral annual crops. Wageningen: Pudoc.
22. Quiring, S. M., & Legates, D. R. (2008). Application of ceres-maize
for within-season prediction of rainfed corn yields in Delaware,
USA. Agricultural and Forest Meteorology, 148,964–975.
23. Ritchie, J. T. (1998). Soil water balance and plant stress. In G. Y.
Tsuji, G. Hoogenboom, P. K. Thornton (Eds.), Understanding op-
tions for agricultural production. Kluwer Academic, Dordrecht, pp.
41/54.
24. Roeckner, A., Brokopf, R., Esch, M., et al. (2003). The atmospheric
general circulation model ECHAM5. Part I: model description. MPI
Report 349. Hamburg: Max Planck Institute for Meteorology. 127 pp.
25. Roeckner, E., Arpe, K., Bengtsson, L., Brinkop, S., Dümenil, L.,
Esch, M., Kirk, E., Lunkeit, F., Ponater, M., Rockel, B., Suasen, R.,
Schlese, U., Schubert, S. and Windelband, M. (1992). Simulation of
the present-day climate with the ECHAM4 model: impact of model
physics and resolution. Max-Planck Institute for Meteorology,
Report No. 93, Hamburg, Germany, 171pp.
26. Roeckner, E., Arpe, K., Bengtsson, L., Christoph, M., Claussen,
M., Dümenil, L., Esch, M., Giorgetta, M., Schlese, U. and
Schulzweida, U. (1996). The atmospheric general circulation model
ECHAM-4: model description and simulation of present-day cli-
mate. Max-Planck Institute for Meteorology, Report No. 218,
Hamburg, Germany, 90pp.
27. Satapathy, S. S., Swain, D. K., & Herath, S. (2014). Field experi-
ments and simulation to evaluate rice cultivar adaptationto elevated
carbon dioxide and temperature in sub-tropical India. European
Journal of Agronomy, 54,21–33.
28. Stocker, T. F., Qin, D., Plattner, G. K., Tignor, M., Allen, S. K.,
Boschung, J., & Midgley, B. M. (2013). IPCC, 2013: climate
change 2013: the physical science basis. Contribution of working
group I to the fifth assessment report of the intergovernmental panel
on climate change.
29. Sudharsan, D., Adinarayana, J., Reddy, D. R., Sreenivas, G.,
Ninomiya, S., Hirafuji, M., & Merchant, S. N. (2013). Evaluation
of weather-based rice yield models in India. International Journal
of Biometeorology, 57,107–123.
30. Warren, R. (2011). The role of interactions in a world implementing
adaptation and mitigation solutions to climate change.
Philosophical Transactions of the Royal Society A: Mathematical,
Physical and Engineering Sciences, 369,217–241.
31. Wilby, R.L., Troni, J., Biot, Y., Tedd, L., Hewitson, B. C., Smith, D.
M., & Sutton,R. T. (2009). A review of climate risk informationfor
adaptation and development planning. International Journal of
Climatology, 29, 1193–1215.
32. Willmott, C. J.,Ackleson, S. G., Davis, R. E., Feddema, J. J., Klink,
K. M., Legates, D. R., O’Donnell, J., & Rowe, C. M. (1985).
Statistics for the evaluation of model performance. Journal of
Geophysical Research, 90(C5), 8995–9005.
33. Willmott, C. J. (1982). Some comments on the evaluation of model
performance. Bulletin of the American Meteorological Society, 63,
1309–1313.
34. Yang, L., Liu, H., Wang, Y., Zhu, J., Huang, J., Liu, G., Dong, G., &
Wang, Y. (2009). Impact ofelevated CO
2
concentration on inter-sub
specific hybrid rice cultivar Liangyoupeijiu under fully open-air
field conditions. Field Crops Research, 112,7–15.
A.V.M.S. Rao et al.
Author's personal copy
