ArticlePDF Available

Abstract and Figures

The complexity of genotype × environment interactions under drought reduces heritability, which determines the effectiveness of selection for drought tolerance and development of drought tolerant varieties. Genetic progress measured through changes in yield performance over time is important in determining the efficiency of breeding programmes in which test cultivars are replaced each year on the assumption that the new cultivars will surpass the older cultivars. The goal of our study was to determine the annual rate of genetic gain for rice grain yield in a drought-prone rainfed system in a series of multi-environment trials conducted from 2005 to 2014 under the Drought Breeding Network of Indian sites in collaboration with the International Rice Research Institute (IRRI). Our results show a positive trend in grain yield with an annual genetic yield increase of about 0.68 % under irrigated control, 0.87 % under moderate reproductive stage drought stress and 1.9 % under severe reproductive stage drought stress due to breeding efforts. The study also demonstrates the effectiveness of direct selection for grain yield under both irrigated control as well as managed drought stress screening to improve yield in typical rainfed systems. IRRI's drought breeding programme has exhibited a significant positive trend in (M.H. Dar), s.hossain@irri.org (S.M. Hossain), a.henry@irri.org (A. Henry), hans-peter.piepho@uni-hohenheim.de (H.P. Piepho). 2 genetic gain for grain yield over the years under both drought stress as well as favorable irrigated control conditions. Several drought tolerant varieties released from the programme have outperformed the currently grown varieties under varied conditions in the rainfed environments on farmers' fields.
Content may be subject to copyright.
Field Crops Research 260 (2021) 107977
0378-4290/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Genetic gain for rice yield in rainfed environments in India
Arvind Kumar
a
,
b
,
*, Anitha Raman
a
, Shailesh Yadav
a
, S.B. Verulkar
c
, N.P. Mandal
d
, O.
N. Singh
e
, P. Swain
e
, T. Ram
e
, Jyothi Badri
f
, J.L. Dwivedi
g
, S.P. Das
h
, S.K. Singh
i
, S.P. Singh
j
,
Santosh Kumar
k
, Abhinav Jain
a
,
l
, R. Chandrababu
m
, S. Robin
m
, H.E. Shashidhar
n
,
S. Hittalmani
n
, P. Satyanarayana
o
, Challa Venkateshwarlu
p
, Janaki Ramayya
p
, Shilpa Naik
p
,
Swati Nayak
q
, Manzoor H. Dar
q
, S.M. Hossain
r
, Amelia Henry
a
, H.P. Piepho
s
a
International Rice Research Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines
b
IRRI South Asia Regional Center (ISARC), Varanasi, India
c
Indira Gandhi Krishi Vishwavidyalaya (IGKV), Raipur, India
d
Central Rainfed Upland Rice Research Station (CRURRS), ICAR-NRRI, Hazaribagh, India
e
National Rice Research Institute (NRRI), Cuttack, India
f
Indian Institute of Rice Research (IIRR), Hyderabad, India
g
Narendra Dev University of Agriculture and Technology (NDUAT), Ayodhya, India
h
ICAR Research Complex for NEH Region, Tripura Centre, Lembucherra, India
i
Banaras Hindu University (BHU), Varanasi, India
j
Bihar Agricultural University (BAU), Sabour, India
k
ICAR-Research Complex for Eastern Region, Patna, India
l
Barwale Foundation, Hyderabad, India
m
Tamil Nadu Agricultural University (TNAU), Coimbatore, India
n
University of Agricultural Sciences (UAS), Bangalore, India
o
Regional Rice Research Station, Maruteru, Acharya NG Ranga Agricultural University, (ANGRAU), Guntur, India
p
International Rice Research Institute, South Asia Hub, ICRISAT, Patancheru, Hyderabad, India
q
International Rice Research Institute (IRRI), New Delhi, India
r
International Rice Research Institute (IRRI), Bhubaneshwar, India
s
Universitaet Hohenheim, Biostatistics Unit, 70593, Stuttgart, Germany
ARTICLE INFO
Keywords:
Rice
Drought
Genetic gain
Yield potential
ABSTRACT
The complexity of genotype ×environment interactions under drought reduces heritability, which determines
the effectiveness of selection for drought tolerance and development of drought tolerant varieties. Genetic
progress measured through changes in yield performance over time is important in determining the efciency of
breeding programmes in which test cultivars are replaced each year on the assumption that the new cultivars will
surpass the older cultivars. The goal of our study was to determine the annual rate of genetic gain for rice grain
yield in a drought-prone rainfed system in a series of multi-environment trials conducted from 2005 to 2014
under the Drought Breeding Network of Indian sites in collaboration with the International Rice Research
Institute (IRRI). Our results show a positive trend in grain yield with an annual genetic yield increase of about
0.68 % under irrigated control, 0.87 % under moderate reproductive stage drought stress and 1.9 % under severe
reproductive stage drought stress due to breeding efforts. The study also demonstrates the effectiveness of direct
selection for grain yield under both irrigated control as well as managed drought stress screening to improve
yield in typical rainfed systems. IRRIs drought breeding programme has exhibited a signicant positive trend in
* Corresponding author at: IRRI South Asia Regional Centre (ISARC), NSRTC Campus, G.T. Road, Collectry Farm P.O. Industrial Estate, Varanasi, 221006, Uttar
Pradesh, India.
E-mail addresses: a.kumar@irri.org (A. Kumar), a.raman@irri.org (A. Raman), shailesh.yadav@irri.org (S. Yadav), satishverulkar@gmail.com (S.B. Verulkar),
npmandal@hotmail.com (N.P. Mandal), onsingh01@yahoo.com (O.N. Singh), pswaincrri@gmail.com (P. Swain), t.ram2011@yahoo.com (T. Ram), jyothirishik@
gmail.com (J. Badri), dwivedi_jl@rediffmail.com (J.L. Dwivedi), drspdas@gmail.com (S.P. Das), shravanbhu1964@gmail.com (S.K. Singh), sps2007bau2011@
gmail.com (S.P. Singh), santosh9239@gmail.com (S. Kumar), a.jain@irri.org (A. Jain), chandrarc2000@yahoo.com (R. Chandrababu), robin.tnau@gmail.com
(S. Robin), heshashidhar@rediffmail.com (H.E. Shashidhar), shailajah_maslab@rediffmail.com (S. Hittalmani), satya_rice@yahoo.com (P. Satyanarayana), c.
venkateshwarlu@irri.org (C. Venkateshwarlu), janakibiotech007@gmail.com (J. Ramayya), shilpa0643@gmail.com (S. Naik), s.nayak@irri.org (S. Nayak),
seedmanzoor.dar@gmail.com (M.H. Dar), s.hossain@irri.org (S.M. Hossain), a.henry@irri.org (A. Henry), hans-peter.piepho@uni-hohenheim.de (H.P. Piepho).
Contents lists available at ScienceDirect
Field Crops Research
journal homepage: www.elsevier.com/locate/fcr
https://doi.org/10.1016/j.fcr.2020.107977
Received 9 December 2019; Received in revised form 1 October 2020; Accepted 5 October 2020
Field Crops Research 260 (2021) 107977
2
genetic gain for grain yield over the years under both drought stress as well as favorable irrigated control
conditions. Several drought tolerant varieties released from the programme have outperformed the currently
grown varieties under varied conditions in the rainfed environments on farmerselds.
1. Introduction
Rice is the most important crop in the world feeding more people
than any other crop, and plays a vital role in the Asian economy. In
1966, the release of IR8, the rst semi-dwarf high yielding modern rice
variety from a cross between Peta, a tall Indonesian variety, and Dee-
geo-woo-gen, a short statured variety from Taiwan, marked the initia-
tion of the green revolution in rice production that over time has
transformed the food decient Asia to a food self-sufcient region. The
green revolution had a remarkable impact on rice production, following
which rice farming underwent signicant transformation. The improved
short-statured varieties (such as Jaya, IR20, IR36, IR42, IR50) developed
since then for the irrigated ecosystem had high yield potential, short
growth duration, were input responsive, and were disease and insect
resistant, however they lacked improved grain quality (Khush, 1987;
Khush and Virk, 2005). IR64, released in the Philippines in 1985 had
superior grain quality as well as higher yield compared with previously
released IR varieties, and its long persistence in farmers elds after its
release was attributed to the excellent eating quality (Champagne et al.,
2010; Mackill and Khush, 2018). The green revolution was based upon
the philosophy that selection can be done under optimal environments
assuming that an increased yield potential will have a carryover effect
under water shortage conditions. However, in the regions chronically
affected by drought, that approach failed to improve yield under
drought conditions (Ceccarelli, 2010). Thus, by focusing on technologies
for very favorable, usually irrigated environments, millions of hectares
of land in rainfed areas were not able to fully benet from yield gains
made in irrigated areas (Ziegler, 1998).
Rice production systems can be classied into lowland and upland.
Rainfed rice elds that are not irrigated but in which the soil is ooded
for at least part of the crop cycle are commonly known as rainfed low-
lands. Water availability is unpredictable, as the crop is rainfed. Asia has
about 46 million hectares of rainfed lowland rice or almost 30 % of the
total world rice area (Maclean et al., 2002). One-third of South and
South East Asian rice lands lie within this ecosystem, which dominates
rice areas of Bangladesh, Cambodia, Myanmar, Nepal, and Thailand,
and is important in India, Indonesia, Laos, and Vietnam (CGIAR, 1998).
More than 75 % of the regions poor rice farmers depend on rainfed
agriculture. The uncertain water supply in rainfed lowland areas,
together with the infertile soil that can be acidic or saline and the
varying crop management methodologies, provide a highly heteroge-
neous set of breeding targets with a range of environmental conditions
that inuence the phenotypic response of the genotypes. This complex
and diverse situation very often lies within small geographical regions.
Genetic improvement under such complex situations would be a key
challenge for any plant breeding program.
In the rainfed environments characterized by lack of water control,
drought and ooding are regular problems. Drought is the most signif-
icant constraint affecting rice production in rainfed lowland rice. The
intensity, duration and timing of drought may vary from location to
location and in a given location from year to year. At reproductive stage,
drought causes a reduction in the number of grains per panicle, increases
grain sterility, and reduces grain weight (Pantuwan et al., 2002).
Drought during the reproductive stage leads to major yield reduction
(Kumar et al., 2007). Even moderate drought stress at reproductive stage
can result in substantial reduction in grain yield (OToole, 1982; Huke
and Huke, 1997; Zhang et al., 2018). Moreover, drought seldom occurs
in isolation; it often interacts with other abiotic and biotic stresses
(Ceccarelli and Grando, 1996) such as soil texture, pH, soil fertility,
diseases and insects. With the effect of climate change, the recurrence
and intensity of drought is likely to change (Ceccarelli et al., 2004) and
possibly becomes more frequent (World Bank, 2009). Recurrent drought
linked with climate change would have an adverse effect on crop pro-
ductivity and on millions of poor farmers livelihood. Therefore,
drought-prone rice systems require stress tolerant rice varieties and
improved management strategies.
However, breeding for drought resistance has been slow. In addition
to drought tolerance being a complex trait, it is reported to be controlled
by several genes each with small effects, and signicant genoty-
pe ×environment interactions (G ×E) under stress (Dixit et al., 2014)
lead to variable performance of genotypes across locations. Conse-
quently, yield under drought has been reported to have low heritability
(H) which complicates selection of superior, drought tolerant, stable
genotypes. In turn, this limits genetic gain for grain yield under drought.
Genetic gain, although a less commonly used measure of breeding ef-
ciency (Ceccarelli, 2015), is dened as the amount of increase in per-
formance that is achieved annually through articial selection (Xu et al.,
2017). Genetic gain estimation is vital for any crop breeding program to
analyze its strengths and weaknesses and to plan future breeding ac-
tivities. In stress environments, improving selection efciency and
increasing genetic gain requires a proper understanding of the target
environment, a selection environment that is representative of the target
environment, a population with large genetic variance, proper trial
management, and appropriate screening procedures.
Genetic gain is usually estimated using multi-environment trials
(METs) that are routinely conducted as a part of breeding program
(Muralidharan et al., 2002; Fufa et al., 2005). Traditionally, popular
check cultivars are grown in all trials in METs to minimize the inuences
of location and year. Genetic progress has been estimated from the
difference between checks and top-yielding cultivars (Feyerherm et al.,
1984) or by subtracting the trend line tted to a set of checks common
across many years from the trend line tted to the year-wise means
(Sharma et al., 2012). However, choice of a representative check
cultivar becomes complicated and with the introduction of new crop
genotypes, the check cultivar may become unrepresentative and a new
check cultivar has to be chosen (Friesen et al., 2016). In previous ana-
lyses, the genotype yields were often expressed as a percentage of the
long-term check yield, assuming that this would minimize G ×E, which
may not be realistic. Moreover, data generated from METs aimed at
screening of new cultivars are highly unbalanced due to variation in the
genotypes tested from year to year, variation in the number of replicates
at each site and variation in the sites that are included from year to year.
In addition, stress-susceptible varieties may contribute to missing values
when exposed to severe stress (Raman et al., 2012). Therefore, the
analytical methods for METs need to be adaptable to unbalanced data
(Liu et al., 2015). Mixed model analysis is an alternative to the tradi-
tional MET analysis based on checks. Missing values, parameter esti-
mation and prediction of genotype performances are effectively handled
by mixed models (Piepho and M¨
ohring, 2006). Treating genotypes as
random effect in a mixed model allows their genetic value to be pre-
dicted using the best linear unbiased predictors (BLUPs). Recently,
mixed models have been used to dissect genetic and non-genetic trends
in multi-environment trial data (Mackay et al., 2011; Piepho et al., 2017;
Laidig et al., 2014, 2017; Ahrends et al., 2018). The contribution of the
cultivars (genetic trend) and of the environments (non-genetic trends) to
yield improvement over time is quantied by regression coefcients.
The genotypes screened in this study are breeding lines developed
from crosses involving a high yielding, but drought susceptible recipient
parent that possesses good grain quality, and a low yielding but highly
drought-tolerant donor parent. The lines were advanced through direct
A. Kumar et al.
Field Crops Research 260 (2021) 107977
3
selection for grain yield (Kumar et al., 2007) under both favorable
conditions as well as variable levels of reproductive-stage drought stress
in different generations in a pedigree generation advancement breeding
program. The selected lines were evaluated for their performance for
grain yield under favorable irrigated condition as well as moderate and
severe levels of reproductive stage drought stresses in the years
20052014 (Verulkar et al., 2010; Mandal et al., 2010). The experiments
were conducted in the Indian sites in collaboration with the Interna-
tional Rice Research Institute (IRRI) under the Drought Breeding
Network of the STRASA (Stress-Tolerant Rice for Africa and South Asia)
project. The STRASA project, supported by the Bill and Melinda Gates
Foundation, has helped millions of farmers who produce their crop
under predominantly rainfed conditions to achieve signicantly higher
yield despite abiotic stresses such as drought, ood, cold, salinity and
iron toxicity (Ismail et al., 2013; Dar et al., 2014; Arora et al., 2019).
The objective of the present study under STRASA was to assess and
monitor yield trends and to estimate the annual rate of increase in yield
for breeding lines in the highly variable drought prone rainfed lowland
environments of eastern and southern India. The trends were dissected
to quantify the contributions of varietal selection and crop management
to the improvement in grain yield. Comparisons draw on yield perfor-
mance and/or tolerance to different biotic/abiotic stresses, of the
released stress tolerant rice varieties (STRVs) as well as the popular
farmer varieties.
2. Material and methods
2.1. Field sites, genotypes and years of screen
The data comprise yield measurements from rainfed lowland trials
conducted under three pre-dened conditions of water and agronomic
management; 1. Irrigated control (maintaining the plants under well-
watered conditions until physiological maturity) conditions, 2. rainfed
with supplementary irrigation and, 3. rainfed conditions.
Approximately 5070 selected genotypes of 100120 days duration
were pooled to constitute an advanced yield trial (AYT). These AYT were
conducted during wet seasons for ten years from 2005 to 2014 at four-
teen on station test sites in India (Table 1,2).
All trials were laid out in alpha-lattice designs with three replica-
tions. Wet-season trials followed a standardized drought-screening
protocol at all sites (Verulkar et al., 2010; Kumar et al., 2014). In
brief, the stress was imposed by draining out water from the eld at 50
DAS (days after seeding) and the cyclic soil moisture decit stress was
maintained till the maturity stage. Perforated PVC pipes of 1.1 m length
was installed at regular places across the eld to measure depth of water
table. A life-saving irrigation was provided as soon as the water table
level in the PVC fell below 1.0 m from soil surface and all the susceptible
checks started showing severe leaf rolling and dying at 10 am. Water was
drained out immediately after 24 h to initiate the next cycle of stress.
The irrigated control trial was planted at each of the sites with
similar crop management practices to determine grain yield reduction
under stress. Standing water was maintained in the eld of the irrigated
control trials from transplanting until maturity. Under reproductive
stage drought, the seeding and transplanting for stress trials was delayed
by approximately 2025 days to increase the probability of the repro-
ductive stage of the crop to coincide with the end of monsoon rains. A
detailed description for a list of typical planting delays at each site was
given previously by Singh et al. (2017).
Most stress trials were conducted in light-soiled elds high in top-
osequence with the crop fully irrigated for one month after transplanting
and then drained to ensure rapid drying. The rainfed condition trials
were planted normally but were never irrigated, and unbunded elds
were chosen to increase the probability of water stress. All three con-
ditions were judiciously designed to reect and simulate conditions of
farmerselds.
To eliminate the bias of border effects, we created our eld layout in
such a way that the planted area was continuous, without leaving space
between plots. The two hills (borders) neighboring the bund (edge of the
eld) were excluded from the harvest for yield assessment in all plots.
All the hills in the plot area described above were harvested, and we
discarded measurements made on plots with more than 20 % missing
hills. A recommended agronomical practice previously reported by
Table 1
Salient features of the STRASA network sites.
Sites Location Target environment Soil type Field topography Drought Presently grown varieties
Indira Gandhi Krishi
Vishwavidyalaya (IGKV),
Raipur, Chhattisgarh
2114N,
8138E
Rainfed low land
ecosystem
Clay, clay loam, low
organic carbon
Bunded, shallow
lowland to mid lowland
Reproductive stage
and early season
drought
MTU 1010, IR 64,
Mahamaya, Swarna
Narendra Dev University of
Agriculture and Technology
(NDUAT), Faizabad, Uttar
Pradesh
2647N,
8212E
Rainfed shallow,
mid and deep low
land
Clay, clay loam, low
organic carbon
Bunded, shallow, mid
and deep lowland
Early season
drought and
reproductive stage
Sarjoo 52, Swarna, NDR
97, NDR 359, Baranideep
Central Rainfed Upland Rice
Research Station (CRURRS),
Hazaribag, Jharkhand
2359N,
8225E
Bunded uplands
and rainfed shallow
low land
Acidic in nature, very
poor in fertility, low in
available N and
organic carbon
Highly undulating,
unbunded, bunded
uplands and bunded
shallow lowland
Drought at all
stages of growth
IR 36, IR 64, MTU 1010,
Hazaridhan, Sadabahar,
Birsa 201, Some other land
races
Central Rice Research Institute
(CRRI), Cuttack, Odisha
2028N,
8554E
Rainfed upland and
shallow lowland in
dry season
Sandy loam and clay
loam soil
Bunded, shallow, mid
and deep lowlands
Seedling, and
vegetative stage
Lalat, Swarna,
Varshadhan, Naveen, MTU
1010, Khandagiri,
Vandana,
TamilNadu Agricultural
University (TNAU),
Coimbatore, TamilNadu
1100N,
7700E
Irrigated lowland Clay Bunded, shallow and
mid lowlands
Reproductive stage IR 64, Co. 47
Tamil Nadu Agricultural
University (TNAU),
Paramakudi, TamilNadu
0931N,
7839E
Rainfed upland Clay loam Bunded Upland Reproductive stage PMK 3, ADT 38, Local
races
University of Agricultural
Sciences (UAS), Bangalore,
Karnatakka
1258N,
7738E
Rainfed upland and
shallow lowland
Red, loamy and light Bunded, shallow
lowland
Reproductive stage MTU 1001, IR 64, Jyoti,
MTU 1010, BPT 5204,
Rasi,
Barwale Foundation (BF),
Hyderabad, Telengana
1720N,
7830E
Rainfed Shallow
low land
Clay loam Bunded, shallow
lowland
Reproductive stage Sambha Mahsuri, Swarna,
MTU 1010
Birsa Agricultural University
(BAU), Ranchi, Jharkhand
2317N,
8519E
Rainfed bunded
uplands and
shallow low land
Acidic in nature, very
poor in fertility, low in
available N and
organic carbon
Highly undulating,
bunded uplands and
shallow lowland
Drought at all
stages of growth
Lalat, IR 36, IR 64, MTU
1010, Birsa 201, Some
other land races
A. Kumar et al.
Field Crops Research 260 (2021) 107977
4
Kumar et al. (2014) was followed for irrigated control as well as drought
trials conducted at IRRI, SA-Hyderabad. In brief, the row-to-row and
plant-to-plant spacing was maintained at 20 cm ×20 cm. Nitrogen,
phosphorus and potassium (NPK) were applied at the rate of
120:60:40 kg ha
1
. N was applied in three doses, rst as basal, second at
tillering stage and third at panicle initiation while P and K were applied
as basal in a single dose. In order to control weeds, Sot (pretila-
chlor ±safene, 0.3 kg a.i. ha
1
) was sprayed just after transplanting
followed by Furadan (carbofuran 1 kg a.i. ha
1
) and Dimotrin (cartap
hydrochloride, 0.25 kg a.i. ha
1
) at 5 and 16 days after transplanting,
respectively. The detailed information on the trials conducted at various
sites in India (location, year of conduct, harvested area, dose of fertil-
izers applied, rainfall during the reproductive-stage stress period) is
presented in Supplementary Table S1.
The experimental sites included in this study represented the major
drought-prone rice-growing regions in India. A description of the land-
use history, soil and environmental characteristics of these sites,
including the 20122014 AYT trials in the present study, was reported
by Singh et al. (2017).
After harvest, each stress trial was classied as moderately stressed if
the yield reduction as compared to the irrigated control trial was 3065
%, and severely drought-stressed if the yield reduction was more than 65
% (Table 2). Trials with less than 30 % yield reduction were classied as
mild stress and were not included in our analysis, since mild drought
stress did not clearly distinguish the lines with high yield had potential
from drought tolerant lines (Verulkar et al., 2010). Thus, the irrigated
control category has 54 trials, 28 trials were classied as moderately
stressed and 31 trials were categorized as severely stressed.
Table 2
Details of the Advanced Yield Trials (AYT) trials assigned to two stress levels based on the reduction in trial mean yield as compared to irrigated control during 2005
2014.
Irrigated control Moderate stress Severe Stress
Trial Location year Trial mean
yield
Trial
H
Trial Location Year Trial mean
yield
Trial
H
Trial Location Year Trial mean
yield
Trial
H
1 Faizabad 2005 6951 0.95 1 Raipur 2006 2796 0.94 1 Faizabad 2006 1265 0.94
2 Hazaribag 2005 4997 0.87 2 Faizabad 2007 2192 0.96 2 Hazaribag 2006 994 0.86
3 Raipur 2005 3814 0.87 3 Hazaribag 2007 3383 0.83 3 Raipur 2006 1223 0.91
4 Faizabad 2006 3658 0.98 4 Raipur 2007 3270 0.77 4 Raipur 2007 1889 0.8
5 Hazaribag 2006 2998 0.51 5 Barwale 2008 2013 0.92 5 TNAU 2007 1025 0.32
6 Raipur 2006 4750 0.61 6 Hazaribag 2008 2234 0.85 6 Barwale 2008 991 0.92
7 Barwale 2007 5480 0.48 7 Raipur 2008 2226 0.84 7 DRR 2008 1004 0.86
8 Faizabad 2007 4098 0.97 8 Faizabad 2009 2369 0.87 8 Faizabad 2008 1639 0.98
9 Hazaribag 2007 4567 0.72 9 Hazaribag 2009 2222 0.83 9 Hazaribag 2008 1605 0.91
10 Raipur 2007 3634 0.73 10 Raipur 2009 2470 0.68 10 Raipur 2008 1009 0.93
11 Barwale 2008 4108 0.84 11 Tripura 2009 2367 0.39 11 TNAU 2008 1126 0.74
12 Faizabad 2008 3183 0.99 12 Barwale 2010 2340 0.72 12 Barwale 2009 1178 0.96
13 Hazaribag 2008 5194 0.84 13 Raipur 2010 2708 0.67 13 Paramakudi 2009 1499 0.46
14 Raipur 2008 3929 0.84 14 TNAU 2010 3711 0.61 14 Raipur 2009 1895 0.7
15 Ranchi 2008 5268 0.73 15 Hazaribag 2011 2507 0.82 15 TNAU 2009 1935 0.98
16 Barwale 2009 3618 0.96 16 Patna 2011 2865 0.92 16 DRR 2010 833 0.96
17 Faizabad 2009 3940 0.97 17 Raipur 2011 2708 0.93 17 Ranchi 2010 1478 0.94
18 Hazaribag 2009 4475 0.93 18 Ranchi 2011 2482 0.7 18 TNAU 2010 1415 0.94
19 Raipur 2009 5258 0.68 19 Barwale 2012 3537 0.82 19 DRR 2011 1341 0.71
20 UAS 2009 4648 0.94 20 Patna 2012 2997 0.93 20 Raipur 2011 1786 0.69
21 Barwale 2010 3904 0.9 21 Tripura 2012 3514 0.98 21 Rewa 2011 1797 0.89
22 Faizabad 2010 3925 0.9 22 DRR 2013 3243 0.87 22 TNAU 2011 1635 0.65
23 Hazaribag 2010 4326 0.9 23 Faizabad 2013 3783 0.91 23 Hazaribag 2012 1815 0.94
24 Raipur 2010 5610 0.81 24 Patna 2013 3436 0.82 24 Rewa 2012 1714 0.59
25 Ranchi 2010 3405 0.97 25 Raipur 2013 3294 0.91 25 Hazaribag 2013 1537 0.66
26 TNAU 2010 4047 0.5 26 Hazaribag 2014 2953 0.89 26 Ranchi 2013 1850 0.55
27 DRR 2011 5371 0.45 27 Raipur 2014 2455 0.93 27 Sabour 2013 1438 0.97
28 Faizabad 2011 4101 0.96 28 Ranchi 2014 2523 0.74 28 TNAU 2013 528 0.97
29 Hazaribag 2011 5488 0.8 29 DRR 2014 958 0.9
30 Patna 2011 4093 0.93 30 Faizabad 2014 602 0.96
31 Raipur 2011 5185 0.78 31 Tripura 2014 1747 0.61
32 Barwale 2012 6035 0.57
33 Faizabad 2012 4445 0.92
34 Hazaribag 2012 5252 0.94
35 Patna 2012 5918 0.94
36 Raipur 2012 6183 0.82
37 Sabour 2012 5135 0.77
38 TNAU 2012 4609 0.73
39 Tripura 2012 5620 0.7
40 DRR 2013 5453 0.89
41 Faizabad 2013 5061 0.87
42 Hazaribag 2013 4804 0.74
43 Patna 2013 5745 0.7
44 Raipur 2013 5151 0.96
45 TNAU 2013 4318 0.84
46 DRR 2014 5645 0.69
47 Faizabad 2014 4087 0.92
48 Hazaribag 2014 6002 0.87
49 IRRI-SA 2014 8336 0.36
50 Maruteru 2014 4451 0.98
51 Raipur 2014 6382 0.92
52 Sabour 2014 4775 0.6
53 Tripura 2014 5896 0.63
54 Varanasi 2014 8284 0.79
A. Kumar et al.
Field Crops Research 260 (2021) 107977
5
The genotypes screened included: (i) advanced breeding lines of
medium (100120 days) duration generated using crosses of popular
high yielding varieties with a diverse array of donors for drought
tolerance, (ii) popular varieties as check, and (iii) traditional drought
tolerant landraces. Most of the popular checks and landraces used in this
study were medium duration, similar to developed breeding lines.
However, some checks like Swarna are long duration in maturity and
were planted 10 days earlier to synchronize owering of breeding lines
in the reproductive-stage. Each year, newly nominated lines from each
participating center and IRRI were jointly evaluated in an augmented-
design in observational yield trials (OYTs), both in irrigated control
conditions and under severe reproductive-stage stress, along with the
checks. Promising lines from OYT screens were promoted to the AYTs
(Verulkar et al., 2010). The criteria for selection of genotypes was based
on grain yield advantage of the genotype shown under irrigated control
as well as under moderate and severe drought stress conditions. The
details of the number of genotypes and their connectivity across years
are presented in Table 3.
2.2. Dissecting genetic and non-genetic trends
The analysis was based on the model dissecting genetic trends from
non-genetic trends in a two-stage analysis provided by Piepho et al.
(2014). Genetic trends are due to breeding efforts, which can be assessed
based on the year a genotype rst entered the trials (birth). Non-genetic
trends are due to agronomic efforts or the external factors that equally
affect all genotypes (trial year), and they can be assessed based on cal-
endar years.
The trials (site ×year combinations) were analyzed using a two-
stage analysis. In the rst stage, data from individual trials was
analyzed using a mixed model that considered genotypes as xed and
replicate and block effects as random to calculate the adjusted means for
the genotypes in the eth trial with an estimated variance-covariance
matrix Ve. If the adjusted means are sorted by trials (by site, year) and
then by genotypes within each trial, V is of the block-diagonal form
V= ⊕m
e=1Ve=
V10 … 0
0V2… 0
⋮ ⋱
0 0 Vm
where is the direct sum operator, m is the number of trials and V
e
is the
estimated variance-covariance matrix of adjusted means in the eth trial.
A weighting matrix is given by DV1
e, a diagonal matrix with diagonal
elements equal to those of V1
e (Smith et al., 2001 and 2005).
In the second stage, for a given stress level, the basic model for the
adjusted means is given by
Yijk =
μ
+gi+yk+ (gy)ik +lj+ (ly)jk + (gl)ij + (gyl)ijk +
ε
ijk − − − − −
>
(1)
where
Yijk is the adjusted mean of the ith genotype in the jth site and kth
year,
μ
is the overall mean, gi is the main effect of the ith genotype, lj is
the main effect of the jth site, yk is the main effect of the kth year, (ly)jk is
the interaction of the jth site with the kth year, (gy)ik is the interaction of
the ith genotype with the kth year, (gl)ij is the interaction of the ith
genotype with the jth site, (gyl)ijk is the interaction of the ith genotype
with the jth site in the kth year and
ε
ijk is the error. The regression terms
for the time trends for the main effects of the genotype effect gi and the
year effect yk are then incorporated into the model. The genotypic main
effect gi is modeled as a function of the year of rst testing (ri) for the ith
variety
gi=βri+Hi
where ,β is the xed regression coefcient for the genetictrend and Hi
is the random deviation of from the genetic trend line, HiN0,
σ
2
H
The year main effect can be modeled using the calendar year t
k
of the kth
year as
yk=γtk+Zk
where γ is the xed regression coefcient for the agronomictrend and
Zk is the random deviation from the agronomic trend line, ZkN0,
σ
2
Z.The model was tted using PROC HPMIXED of SAS (SAS 9.3).
Damesa et al. (2017) provided a SAS macro to calculate the weights
proposed by Smith et al. (2001) for the two-stage analysis. Percentage
change in genetic gain per year for each stress level due to the genetic
and non-genetic causes were estimated as;
reg.coeff
μ
+ (reg.coeff x start year mean)x 100
where
μ
is the overall intercept (the grand mean across all the groups
with no predictors), regression coefcient is β and γ for genetic and non-
genetic causes, respectively and start year is start of the period over
which gain was assessed.
2.3. Head-to-head trials
Head-to-head trials are a method of eld demonstration initiated in
the STRASA project, which allow the farmers to assess the yield per-
formance of new variety with the already existing one under variable
growing conditions in a rainfed system. The STRVs developed under
STRASA were evaluated in head-to-head trials in farmerselds in the
eastern Indian state of Odisha under two ecosystems: (i) the lowland and
(ii) mid -and -shallow lowland. In the current study, within the lowland
ecosystem, in addition to the topography, areas prone to submergence
are referred to lowland, areas prone to occurrence of both drought and/
or submergence in the same or different seasons have been referred to
mid lowland and areas prone to frequent drought stress are referred to
shallow lowland.
In the head-to-head trials, the pair of varieties to be compared were
grown in the same eld (half of the eld for each variety) or in two
adjacent elds with similar conditions. The head-to-head trials were
conducted using the farmers management practices in the particular
locations to assess the extent of yield gain achieved on the farmerselds
(Dar et al., 2018). The data was analyzed separately for the lowland and
shallow-and-medium lowland ecosystems.
The data was subjected to analysis of variance with xed effects for
variety, and random effects for districts and farmers nested within dis-
tricts. Yield of the ith variety from the kth farmer in the jth district is
Table 3
Details of the number of genotypes and their connectivity across years under
irrigated control and the two stress levels in AYT trials.
No. of years tested
No. of genotypes
Irrigated control Moderate
stress
Severe
stress
1 242 214 214
2 72 128 58
3 30 69 19
4 15 64 14
5 3 5 14
6 1 12 3
7 1 8
8 9 1
9 1 1
10 1
Total no. of genotypes 366 325 324
A. Kumar et al.
Field Crops Research 260 (2021) 107977
6
modelled by
yijk =
μ
+ti+ej+fjk + (te)ij +
ε
ijk (2)
where,
μ
is the overall mean, ti is the effect of the ith variety, ej is the
effect of the jth district, fjk is the effect of the kth farmer in the jth dis-
trict, teij is the effect of the district by variety interaction,
ε
ijk is the error.
Heterogeneity of error variances among environments was investigated
by examining a plot of the standardized residuals against the tted
response for both models (Raman et al., 2012; Hu et al., 2013). The
models with homogeneous and heterogeneous error variances for dis-
tricts were compared. The best model was chosen depending on the AIC
values. In each ecosystem, the individual varieties were then classied
into (i) STRVs and (ii) farmers variety. The model was rerun to compare
the performances of the two categories.
3. Results
3.1. Single trial analysis of AYT
The mean grain yield ranged from 2.58.0 t/ha in the irrigated
control trials, 23.7 t/ha in the moderate drought stress trials and
0.51.9 t/ha under severe drought stress trials. The trials with herita-
bility less than 0.30 were excluded from the subsequent analysis
(Johnson et al., 1955). This included 7 trials under control, 6 under
moderate stress and 5 under severe stress. Heritability of the trials
included in the analysis ranged from 0.36 to 0.99 under non-stress, 0.39
to 0.98 under moderate stress and 0.32 to 0.98 under severe stress
Fig. 1. Adjusted mean yields from single trial-wise analysis of AYT.
A. Kumar et al.
Field Crops Research 260 (2021) 107977
7
(Table 2). A plot of the trial-wise adjusted means under the irrigated
control and two stress levels is shown in Fig. 1(a, b & c).
3.2. Analysis across years and sites
Comparing the regression estimates for the genetic and the non-
genetic trends, it is observed that the genetic trend was positive with
an increase in yield at the rate of 0.034 t/ha (condence interval, CI:
0.007, 0.060) per year under irrigated control where the numbers in
parenthesis indicate the lower and upper condence limits at
α
=0.05.
The trend under moderate stress was 0.025 t/ha (CI: -0.004, 0.053) and
0.027 t/ha (CI: 0.004, 0.053) under severe stress. This can be inter-
preted as an increase of about 0.68 % p.a. under non-stress, 0.87 % p.a.
under moderate stress and 1.9 % p.a. under severe stress (Table 4). The
non-genetic trend was positive with an increase of about 0.122 t/ha (CI:
0.005, 0.239) under irrigated control and 0.05 t/ha (CI: -0.062, 0.135)
in moderate stress, respectively. Under severe stress, a negative trend
was found with a yield decrease of 0.021 t/ha (CI: -0.083, 0.051)
(Table 4). The percentage increase was about 2.3 % p.a. under non-
stress, 1.89 % p.a. under moderate stress and 1.5 % p.a. decrease
under severe stress. A plot of marginal means of the variety groups based
on the year of rst testing (r
i
) and the calendar year (t
k
) for the irrigated
control and two stress levels are presented in Fig. 2. In all the conditions,
the year ×site interaction variance was relatively large compared to the
variances of all effects involving genotype. The genotype ×year and
genotype ×site interaction variances were small compared to the gen-
otype ×site ×year interaction variance (Table 5).
3.3. Analysis of head-to-head trials
The varietal differences were non-signicant (p <0.05) in WS2017
and signicant in WS2018 in lowland trials while it was highly signi-
cant in both years in the shallow and medium lowland trials based on
model 3 (Supplementary Table S2). Swarna -Sub1 had produced the
highest yield in the lowland ecosystem in both years while DRR dhan 44
had yielded highest in the shallow and medium lowland situation
(Table 6). The STRVs outperformed the currently grown varieties on
farmerselds in all situations in both years (Fig. 3. a,b,c & d).
4. Discussion
The increase in frequency and intensity of drought due to climate
change necessitates the use of varieties that possess (i) drought resis-
tance together with high yield under favorable conditions (Serraj et al.,
2011), (ii) appropriate growth duration based on toposequence of the
land, (iii) resistance to biotic stresses, (iv) improved grain quality and
(v) other traits of economic importance. These goals can be achieved
with large-scale breeding programs targeted to achieve a continuous
cycle of varietal improvement, wherein older varieties are replaced with
new varieties that can meet farmersneeds and thus deliver genetic
gains.
Genetic gain assessment is vital for evaluating genetic improvement,
selection efciency and for designing future breeding strategies (Fufa
et al., 2005). Breeding programs designed for the rainfed ecosystem
targeting the developing world will need to deliver higher rates of ge-
netic gain to cope with the 21 st century challenges of 50 %60 %
greater demands for food commodities, climate change and natural
resource constraints (Genetic Gains Working Group, 2016). IRRI, in
collaboration with national research partners under STRASA, initiated
efforts for developing drought-tolerant rice varieties suitable for rainfed
lowland drought-prone areas of India. Many popular high yielding rice
varieties are highly susceptible to drought, hence IRRI implemented a
breeding strategy for the development and evaluation of breeding lines
in on-station trials under control as well as moderate stress and severe
reproductive stage drought stress (Kumar et al., 2007; Venuprasad et al.,
2007; Verulkar et al., 2010; Mandal et al., 2010) with the aim to
combine drought tolerance with high yield potential. Screening under
moderate and severe drought stress exposed the breeding lines to a
range of severity of reproductive stage drought stresses likely to prevail
in the rainfed regions. The efforts led to development of several high
yielding drought tolerant lines (Kumar et al., 2014). Genetic gain
assessment of grain yield in the present study clearly indicates that the
efforts have been successful in not only increasing the yield under
moderate and severe reproductive stage drought stress but also under
irrigated control. Further, the evaluation of breeding lines at segregating
generations as well as at the advanced stage under the three situations
ensured obtaining genetic gains under all the three conditions. The ge-
netic gain of 0.68 % under control, 0.87 % under moderate stress and 1.9
% under severe reproductive stage drought stress demonstrates the
success of our breeding strategy implemented between 20042014.
Earlier, Xangsayasane et al., 2013 reported that screening of lines under
intermittent drought in the dry season and terminal drought in the wet
season provided the magnitude of drought severity that was appropriate
identifying promising drought tolerant lines. Based on drought response
index (DRI), genotypes shown to be drought tolerant were consistent in
performance in intermittent and terminal drought screening (Xang-
sayasane et al., 2013). Promising genotypes with good yield potential
under irrigated control as well as under moderate and severe drought
stress had been previously selected based on drought yield index (DYI)
and mean yield index (MYI) in rainfed drought prone locations in
Eastern India (Raman et al., 2012).
The genetic gains achieved by the breeding programs in on-station
trials between 20042014 were also observed in the on-farm trials
conducted during 20172018. The performance of some of the devel-
oped breeding lines released as varieties such as Swarna -Sub 1 for
lowland area, Sahbhagi dhan, DRR dhan 42, DRR dhan 44, Bina dhan 11
for mid and shallow lowland areas were evaluated in comparison with
the varieties farmers were growing (Swarna, CR1009, Khandagiri, Lalat,
MTU1010, Naveen, Pooja, Sarala) in the rainfed regions in on-farm
trials. Swarna-Sub1 showed an average yield advantage of 0.315 and
0.510 t/ha in 2017 and 2018 respectively over mean performance of
varieties farmers were growing in the lowland including Swarna. In
Table 4
Estimates of regression coefcients for the genetic and non-genetic trends and
estimates of their group means from the analysis across sites and years for the
AYT trials under the irrigated control, moderate and severe drought stress.
Irrigated control Moderate stress Severe stress
Regression
coefcient
est std. err est std. err est std. err
β 0.034 0.014 0.025 0.015 0.027 0.012
γ 0.122 0.059 0.055 0.055 0.021 0.032
Group means
Year r
i
t
k
r
i
t
k
r
i
t
k
2005 4.793 5.258
2006 4.673 3.861 2.730 2.656 1.384 1.164
2007 4.765 4.476 2.630 3.042 1.270 1.537
2008 4.682 4.398 2.593 2.343 1.279 1.401
2009 4.795 4.331 2.764 2.458 1.404 1.732
2010 5.033 4.119 2.724 2.999 1.507 1.284
2011 4.939 4.681 2.796 2.672 1.434 1.653
2012 5.047 5.212 2.948 3.290 1.787 1.529
2013 4.578 5.148 2.765 3.436 1.544 1.188
2014 4.450 6.271 3.221 2.272 1.195 1.315
average 4.775 4.775 2.797 2.797 1.422 1.422
% gain
#
0.68 2. 3 0.87 1.9 1.9 1.5
Note: est estimate; std. err standard error. r
i
year of rst testing genetic
trend ; t
k
trial year non genetic trend; β is the xed regression coefcient for
the genetictrend r
i
; γ is the xed regression coefcient for the agronomic
trend t
k.
#
Percentage change in genetic gain per year for each stress level due to the
genetic and non-genetic causes were estimated using the formula explained
under subheading dissecting genetic and non-genetic trends in materials and
methods section.
A. Kumar et al.
Field Crops Research 260 (2021) 107977
8
shallow and medium lowland, the drought tolerant varieties such as
Sahbhagi dhan showed an average yield advantage of 0.452 and
0.554 t/ha in 2017 and 2018 respectively over the mean performance of
varieties under cultivation by farmers. These ndings are in agreement
with the Xangsayasane et al., 2014 who have earlier reported that
drought tolerance determined in either intermittent or terminal drought
in on-station trials was effective in predicting on-farm yield.
For long-term genetic improvement of yield, evaluating the genetic
variation present in the germplasm is a pre-requisite (Allard, 1960;
Falconer and Mackay, 2004), as availability of adequate variability
provides a range of choices from which selections can be made. Genetic
variability for yield under drought stress in rainfed lowland rice in both
dry and wet seasons have been documented (Venuprasad et al., 2007;
Kumar et al., 2007; Verulkar et al., 2010; Gouda et al., 2012; Saikumar
et al., 2016; Shaibu et al., 2017). The genotypic variance component
observed for yield under irrigated control and two stress levels in the
present study was larger than that of the genotype ×year and the gen-
otype ×site variance components suggesting enough genetic variability
in the breeding material used.
When assessing cultivar performance and genetic gains at multiple
sites and across years, breeding and varietal selection may not be the
only major factors responsible the increase in grain yields (Ahrends
et al., 2018; Muralidharan et al., 2019). Long-term genetic gains for
yield under a range of drought situations can be achieved by combining
genetics and crop management (Duvick et al., 2004; Cooper et al.,
2014). This is because responses to drought stress are complex and also
involve climatic, soil and agronomic factors and therefore predictions of
phenotypic response should be made as functions of genetic and envi-
ronment factors. Selection of superior genotypes depends on the quality
of these predictions (Sadras et al., 2013; Bustos-Korts et al., 2016). In the
present study, regression estimates show that in all stress levels, yield
increase is mainly caused by new varieties rather than by the agronomic
Fig. 2. Marginal means of the variety groups based on the year of rst testing (r
i
) and the calendar year (t
k
).
Table 5
REML parameters from the analysis across sites and years for the AYT trials under irrigated control and the two stress levels.
Cov Parm Irrigated control Moderate stress Severe stress
Estimate Standard Error p-value Estimate Standard Error p-value Estimate Standard Error p-value
Year 0.0841 0.1454 0.281 0.112 0.086 0.096 0.004 0.034 0.453
Gen 0.080 0.015 <.0001 0.043 0.013 0.001 0.023 0.008 0.003
Site 0.031 0.141 0.413 0 . . 0.034 0.041 0.206
year*site 0.918 0.231 <.0001 0.132 0.045 0.002 0.147 0.057 0.005
gen*year 0.000 0.010 0.500 0.000 0.011 0.500 0.003 0.007 0.343
gen*site 0.040 0.019 0.025 0.053 0.018 0.002 0.021 0.011 0.025
gen*year*site 0.339 0.028 <.0001 0.162 0.019 <.0001 0.161 0.013 <.0001
Residual 1 0 1 0 1 0
Note: gen genotype; Cov Parm-covariance parameter; REML: restricted maximum likelihood.
A. Kumar et al.
Field Crops Research 260 (2021) 107977
9
factors. The contribution of the environmental factors under irrigated
control and moderate stress were positive while it was negative in the
case of severe stress trials. This could be attributed to the drought stress
in these environments. Assessing the size of the contributions of genetic,
environmental and management factors can have implications for
on-farm planning, testing breeding strategies and allocation of resources
for yield improvement (Anderson, 2010).
Determination of genetic progress in grain yield over a period has
been extensively studied in other grain crops such as wheat and maize
but has been limited in the case of rice. Peng et al. (2000) determined the
trend in rice yield of cultivars developed by IRRI between 1966 and
1996. Their study indicated a gain of 1% per year. Muralidharan et al.
(2002) analysed the grain yield performance of rice genotypes devel-
oped and tested since 19761997 in the METs performed worldwide
under different ecosystems in the international rice improvement trials
and concluded that there was no evidence for either a genetic gain or
loss in grain yields of genotypes developed for any of the ecosystems.
Further, Muralidharan et al. (2019) observed a signicant yield gap
between the projection of human population and national rice grain
production by estimating the genetic gain for yields in genotypes tested
in 11 rice ecosystems from 1995 to 2013 in India. Integrated genomic
technologies and policy interventions is needed to seal this yield gap and
to ensure rice availability to fulll the demands of growing population.
Bastasa et al. (2015) determined the annual genetic gain in rice yield
based on 28 cultivars representing different years of varietal release over
the time period 19662013 and reported an annual yield gain of 1% on
the basis of yield of IR64. Saito (2018) assessed the trend in grain yield
and its associated traits of rice varieties released during the last decades
in sub-Saharan Africa. The study showed considerable genetic gains in
rainfed systems and high yielding lowland varieties released between
1986 and 2013 had around 20 % higher yield than IR64.
In studies assessing the genetic gain for grain yield in different crops,
yield is regressed against the year of release to monitor the gains in
released varieties over time as also reported in other crops (Sayre et al.,
1997; Lopes et al., 2012; Masuka et al., 2017). An alternate method of
assessing long-term genetic gain based on MET data is to evaluate
cultivar performance trials sequentially over a period of time in the
collaborating test sites such as in the present study. Grain yield regressed
against the year of testing and the performance of the test varieties are
compared with that of common checks with the main objective of va-
riety release (Trethowan et al., 2002; Tadesse et al., 2013).
Among the drought tolerant varieties developed and released
following strategy of evaluation of breeding lines for grain yield under
irrigated control and reproductive stress, Sahbhagi dhan, was the rst
variety released in 2009 as a result of the collaboration between IRRI
and different Indian institutions for drought-prone shallow lowland
ecosystems of India (Basu et al., 2017) as well as for bunded uplands.
Sahbhagi dhan had a comparatively stable performance and adaptable
plant types for shallow rainfed lowland drought-prone ecosystems
(Verulkar et al., 2010; Kumar et al., 2012; Raman et al., 2012) and has
shown a yield advantage of 0.81.0 t/ha over other varieties under
drought conditions (Yamano et al., 2013) where the yield of irrigated
control varieties such as IR 64 collapsed (Dar et al., 2014). DRR dhan 44,
released in the year 2014 for cultivation under irrigated conditions, is
characterized by high yield under water limited conditions. Other
drought-tolerant varieties released for commercial cultivation in India
include DRR dhan 42, 43, 50 and CR dhan 204, 205, 801 and 802.
In lowland, Swarna- Sub1 performed better than any other variety
closely followed by CR 1990 Sub 1. In mid lowland, in lower bunds, Bina
dhan 11 performed exceedingly well whereas in mid lowland upper
bunds as well as in shallow lowland DRR dhan 44 performed better than
other varieties including the most presently cultivated variety MTU1010
(Table 6).
5. Conclusion
This study documents the signicant genetic gain for grain yield of a
breeding program targeting rainfed lowland rice in India that was based
on direct selection for grain yield under both irrigated control and
drought conditions. The study utilized extensive multi-season evalua-
tion in target environments under irrigated control, moderate drought
stress and severe drought stress between 20042014 with number of
popular varieties as checks to enable accurate estimation of the genetic
gain. The yield improvement of the newly developed stress-tolerant
varieties over the best currently grown varieties was also demon-
strated on farmers elds. The developed STRVs have potential to pro-
tect farmers from crop losses against an increasing impact of extreme
droughts under climate change. The results of this study shall assist
governments and policy makers to replace the currently grown decades-
old varieties from the seed chain and emphasize larger efforts for seed
multiplication and dissemination of the newly developed stress-tolerant
varieties in coordination with various stakeholders. The ndings of the
Table 6
Estimated variety means from the head-to-head trials at Odisha under lowland and shallow & medium lowland conditions.
Variety type Lowland Shallow & medium lowland
WS2017 WS2018 WS2017 WS2018
Mean std err Mean std err Mean std err Mean std err
Farmer 4.540 0.410 5.818 0.248 3.788 0.266 4.154 0.217
STRV 4.855 0.410 6.328 0.248 4.240 0.266 4.708 0.217
STRV
Bina dhan 11 4.193 0.267 4.621 0.233
Swarna Sub 1 4.873 0.438 6.2157 0.1741
CR 1009 Sub -1 4.846 0.482 6.5717 0.1924
DRR dhan 42 4.203 0.270
DRR dhan 44 4.522 0.276 4.777 0.243
Sahbhagi dhan 4.068 0.286 4.593 0.246
Farmer variety
CR 1009 3.795 0.592 6.2697 0.2094
Khandagiri 3.417 0.423 4.036 0.326
Lalat 3.732 0.270 3.978 0.239
MTU 1001 4.466 0.497 4.253 0.308 4.313 0.258
MTU 1010 3.891 0.312 4.138 0.250
Naveen 3.758 0.267 4.067 0.252
Pooja 4.430 0.478 5.3043 0.1952
Sarala 4.304 0.789 5.7207 0.3094
Swarna 4.693 0.439 6.119 0.1833 3.985 0.386 4.855 0.317
Swarna Shreya 4.870 0.359
Note: STRV-stress tolerant rice varieties.
A. Kumar et al.
Field Crops Research 260 (2021) 107977
10
study shall also encourage researchers to plan for effective evaluation of
the breeding programs products with aim to assess the success of the
breeding programs in terms of genetic gains achieved.
Availability of data and materials
The datasets supporting the conclusions of this article are provided
within the article.
Authorscontribution
AK: was involved in design of overall experiment, drafting and
critical revision of MS; AR: was involved in experimental analysis,
interpretation of data, drafting and revising the manuscript; SY: was
involved in evaluating the lines at IRRI and revising the manuscript;
SBV: conducted trials and recorded measurements at IGKV Raipur;
NPM: performed trials at Hazaribagh, ONS and PS: conducted the trials
and recorded measurements at NRRI Cuttack, TR and JB: conducted the
trial and recorded observations at IIRR-Hyderabad; JLD: performed the
Fig. 3. District-wise predicted means of the STRV and farmer varieties from head to head trials.
A. Kumar et al.
Field Crops Research 260 (2021) 107977
11
trials at NDUAT-Ayodhya; SPD: conducted the trials at ICAR-Tripura,
SKS: conducted the trial at BHU-Varanasi; SPS: conducted the trial
and recorded data at BAU-Sabour; SK: conducted the trial at ICAR-
Patna; AJ: conducted trials and recorded observations at Barwale
foundation, Hyderabad; RC and SR: conducted trials and recorded ob-
servations at TNAU Coimbatore; HES and SH: conducted the trials and
recorded data at UAS-Bangalore, PS: conducted the trials and recorded
the required data at ANGRAU-Guntur, VC, JR and SN: were involved in
multiplication of seeds conducted the trials at IRRI-SAH, Hyderabad,
and distributed the multiplied seeds to partners at various locations; SN,
MD and MH: had monitored and evaluated the trials at various locations
in India; and AH: performed screening of donors for drought tolerance at
IRRI under the STRASA project. HPP: provided suggestions regarding
the experiments and statistical models and analysis and also helped in
editing of the MS. All authors have read and approved the nal
manuscript.
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Acknowledgements
This study was supported by the Bill and Melinda Gates Foundation
(BMGF) through STRASA (Stress Tolerant Rice for Africa and South Asia,
phase 1-3) project. The authors thank BMGF for the nancial support to
the study. The authors are also thankful to the Indian Council of Agri-
cultural Research (ICAR), New Delhi, as well as all the State Agricultural
Universities associated with the study for support provided in con-
ducting the trials at various locations in India. HPP was supported by
DFG grant PI 377/20-1.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.fcr.2020.107977.
References
Ahrends, H.E., Eugster, W., Gaiser, T., Rueda, V., Hüging, H., Ewert, F., Siebert, S., 2018.
Genetic yield gains of winter wheat in Germany over more than 100 years
(18952007) under contrasting fertilizer applications. Environ. Res. Lett. 13,
104003.
Allard, R.W., 1960. Principles of Plant Breeding. John Wiley and Sons Inc, New York,
USA.
Anderson, W., 2010. Closing the gap between actual and potential yield of rainfed wheat.
The impacts of environment, management and cultivar. Field Crops Res. 116, 1422.
Arora, A., Bansal, S., Patrick, S.W., 2019. Do farmers value rice varieties tolerant to
droughts and oods? Evidence from a discrete choice experiment in Odisha, India.
Water Res. and Econ. 25, 2741.
Basu, S., Jongerden, J., Ruivenkamp, G., 2017. Development of the drought tolerant
variety Sahbhagi Dhan: exploring the concepts commons and community building.
Int. J. Commons. 11 (1), 144170.
Bustos-Korts, D., Malosetti, M., Chapman, S., van Eeuwijk, F., 2016. In: Yin, X., Struik, P.
C. (Eds.), Modeling of Genotype by Environment Interaction and Prediction of
Complex Traits across Multiple Environments as a Synthesis of Crop Growth
Modelling, Genetics and Statistics. Springer International Publishing Switzerland. Cr
Systems Biol.
Ceccarelli, S., 2010. Drought and drought resistance. Encyclopedia Biotech. Agri. Food 1,
205207.
Ceccarelli, S., 2015. Efciency of plant breeding. Crop Sci. 55, 8797.
Ceccarelli, S., Grando, S., 1996. Drought as a challenge for the plant breeder. Plant
Growth Regul. 20, 149155.
Ceccarelli, S., Grando, S., Baum, M., Udupa, S.M., 2004. Breeding for drought resistance
in a changing climate. In: Rao, S.C., Ryan, J. (Eds.), In Challenges and Strategies for
Dryland Agriculture. CSSA Special Publication ASA and CSSA, Madison, WI,
pp. 167190. No. 32.
CGIAR, 1998. Press Release ‘CGIAR Urges Halt to Granting of Intellectual Property Rights
for Designated Plant Germplasm. February 11.
Champagne, E.T., Bett-Garber, K.L., Fitzgerald, M.A., Grimm, C.C., Lea, J., Ohtsubo, K.,
Jongdee, S., Xie, L.H., Bassinello, P.Z., Resurreccion, A., Ahmad, R., Habibi, F.,
Reinke, R., 2010. Important sensory properties differentiating premium rice
varieties. Rice 3, 270281.
Cooper, M., Gho, C., Leafgren, R., Tang, T., Messina, C., 2014. Breeding drought-tolerant
maize hybrids for the US corn-belt: discovery to product. J. Exp. Bot. 65, 61916204.
Damesa, T.M., M¨
ohring, J., Worku, M., Piepho, H.P., 2017. One step at a time: stage-wise
analysis of a series of experiments. Agron. J. 109, 845857.
Dar, M.H., Singh, S., Singh, U.S., Ismail, A.M., 2014. Stress tolerant rice varieties: making
headway in India. SATSA Mukhapatra Annu. Tech. Issue 18 (2014), 114.
Dar, M.H., Zaidi, N.W., Waza, S.A., Verulkar, S.B., Ahmed, T., Singh, P.K., Roy, S.B.,
Chaudhary, B., Yadav, R., Islam, M.M., Iftekharuddaula, K.M., 2018. No yield
penalty under favorable conditions paving the way for successful adoption of ood
tolerant rice. Sci. Rep. 8 (1), 9245.
Dixit, S., Singh, S., Kumar, A., 2014. Rice breeding for high grain yield under drought: a
strategic solution to a complex problem. Int. J. Agron. 2014, 112.
Duvick, D.N., Smith, J.S.C., Cooper, M., 2004. Long-term selection in a commercial
hybrid maize breeding program. In: Janick, J. (Ed.), Plant Breeding Reviews: Long-
Term Selection: Crops, Animals, and Bacteria, 24. John Wiley & Sons, New York,
pp. 109151. Part 2.
Falconer, D.S., Mackay, T.F.C., 2004. An introduction to quantitative genetics. Genetics
167 (4), 15291536.
Feyerherm, A., Paulsen, G.M., Sebaugh, J.L., 1984. Contribution of genetic improvement
to recent wheat yield increases in the U.S.A. Agron. J. 76, 985990.
Friesen, L.F., Brûl´
e-Babel, A.L., Crow, G.H., Rothenburger, P.A., 2016. Mixed model and
stability analysis of spring wheat genotype yield evaluation data from Manitoba,
Canada. Can. J. Plant Sci. 96, 305320.
Fufa, H., Baenziger, P., Beecher, B.S., Graybosch, R.A., Eskridge, K.M., Nelson, A.L.,
2005. Genetic improvement trends in agronomic performances and end-use quality
characteristics among hard red winter wheat cultivars in Nebraska. Euphytica 144,
187198.
Genetic Gains Working Group, 2016. CGIAR Platform on Genetic Gains - Tools and
Services to Accelerate Genetic Gains of Breeding Programs Targeting the Developing
World. Proposal to the CGIAR Fund Council, 31 March 2016. International Maize
and Wheat Improvement Center (CIMMYT), Mexico. D.F., MEXICO.
Gouda, P.K., Varma, C.M.K., Saikumar, S., Kiran, B., Shenoy, V., Shashidhar, H.E., 2012.
Direct selection for grain yield under moisture stress in Oryza sativacv. IR58025B
×Oryza meridionalis population. Crop Sci. 52, 644653.
Hu, S.X., Yan, S., Shen, K., 2013. Heterogeneity of error variance and its inuence on
genotype comparison in multi-location trials. Field Crops Res. 149, 322328.
Huke, R.E., Huke, E.H., 1997. Rice Area by Type of Culture, South, Southeast and East
Asia. International Rice Research Institute, Los Banos (Philippines).
Ismail, A.M., Singh, U.S., Singh, S., Dar, M.H., Mackill, D.J., 2013. The contribution of
submergence-tolerant (Sub1) rice varieties to food security in ood-prone rainfed
lowland areas in Asia. Field Crops Res. 152, 8393.
Johnson, H.W., Robinson, H.F., Comstock, R.E., 1955. Estimates of genetic and
environmental variability in soybeans. Agron. J. 47, 314318.
Khush, G.S., 1987. Rice breeding: past, present and future. J. Genet. 66, 195216.
Khush, G.S., Virk, P.S., 2005. IR Varieties and Their Impact. International Rice Research
Institute, Los Ba˜
nos (Philippines), p. 163.
Kumar, A., Venuprasad, R., Atlin, G.N., 2007. Genetic analysis of rainfed lowland rice
drought tolerance under naturally-occurring stress in eastern India: heritability and
QTL effects. Field Crops Res. 103, 4252.
Kumar, A., Verulkar, S.B., Mandal, N.P., et al., 2012. High yielding, drought tolerant,
stable rice genotypes for the shallow rainfed lowland drought prone ecosystem. Field
Crops Res. 133, 3747.
Kumar, A., Dixit, S., Ram, T., Yadaw, R.B., Mishra, K.K., Madal, N.P., 2014. Breeding
high yielding drought-tolerant rice: genetic variations and conventional and
molecular approaches. J. Exp. Bot. 65, 62656278. https://doi.org/10.1093/jxb/
eru363.
Laidig, F., Piepho, H.P., Drobek, T., Meyer, U., 2014. Genetic and non-genetic long-term
trends of 12 different crops in German ofcial variety performance trials and on-
farm yield trends. Theor. Appl. Genet. 127, 25992617.
Laidig, F., Piepho, H.P., Rental, D., Drabik, T., Meyer, U., Huesker, A., 2017. Breeding
progress, environmental variation and correlation of winter wheat yield and quality
traits in German ofcial variety trials and on-farm during 19832014. Theor. Appl.
Genet. 130, 223245.
Liu, W., Mu, C., Ji, R., Shaquan, M., John, R.S., Chang, S., 2015. Low-rank similarity
metric learning in high dimensions. In Proceedings of the Twenty-Ninth AAAI
Conference on Articial Intelligence 27922799 pages.
Lopes, M.S., Reynolds, M.P., Manes, Y., Singh, R.P., Crossa, J., Braun, H.J., 2012. Genetic
yield gains and changes in associated traits of CIMMYT spring bread wheat in
aHistoricset representing 30 years of breeding. Crop Sci. 52, 11231131.
Mackay, I., Horwell, A., Garner, J., White, J., McKee, J., Philpott, H., 2011. Reanalyses of
the historical series of UK variety trials to quantify the contributions of genetic and
environmental factors to trends and variability in yield over time. Theor. Appl.
Genet. 122, 225238.
A. Kumar et al.
Field Crops Research 260 (2021) 107977
12
Mackill, D.J., Khush, G.S., 2018. IR64: a high-quality and high-yielding mega variety.
Rice 11, 18.
Maclean, J.L., Dawe, D., Hardy, B., Hettel, G.P. (Eds.), 2002. Rice Almanac. IRRI,
WARDA, CIAT, FAO, Los Ba˜
nos (Philippines), Bouak´
e (Cˆ
otedIvoire), Cali
(Colombia) and Rome (Italy), p. 253.
Mandal, N.P., Sinha, P.K., Variar, M., Shukla, V.D., Perraju, P., Mehta, A., Pathak, A.R.,
Dwivedi, J.L., Rathi, S.P.S., Bhandarkar, S., Singh, B.N., Singh, D.N., Panda, S.,
Mishra, N.C., Singh, Y.V., Pandya, R., Singh, M.K., Sanger, R.B.S., Bhatt, J.C.,
Sharma, R.K., Raman, A., Kumar, A., Atlin, G., 2010. Implications of genotype x
input interactions in breeding superior genotypes for favorable and unfavorable
rainfed upland environments. Field Crops Res. 118, 135144.
Masuka, B., van Biljon, A., Cairns, J.E., Das, B., Labuschagne, M., MacRobert, J.,
Makumbi, D., Magorokosho, C., Zaman-Allah, M., Ogugo, V.O., Prasanna, B.M.,
Tarekegne, A., Semagn, K., 2017. Genetic diversity among selected elite CIMMYT
maize hybrids in East and Southern Africa. Crop Sci. 57, 23952404.
Muralidharan, K., Prasad, C.S.V., Rao, C.S., 2002. Yield performance of rice genotypes in
international multi-environment trials during 197697. Current Sci. 83, 610619.
Muralidharan, K., Prasad, C.S.V., Rao, C.S., Siddiq, E.A., 2019. Genetic gain for yield in
rice breeding and rice production in India to meet with the demand from increased
human population. Current Sci. 116 (4), 544560.
OToole, J.C., 1982. Adaptation of rice to drought prone environments. Drought
Resistance in Crops with Emphasis on Rice. IRRI, Los Ba˜
nos, Philippines, p. 195.
Pantuwan, G., Fukai, S., Cooper, M., Rajatasereekul, S., OToole, J.C., 2002. Yield
response of rice (Oryza sativa L.) genotypes to drought under rainfed lowland
contributing to drought resistance. Field Crops Res. 73, 181200.
Peng, S., Laza, R.C., Visperas, R.M., Sanico, A.L., Cassman, K.G., Khush, G.S., 2000. Grain
yield of rice culitvars and lines developed in Philippines since 1996. Crop Sci. 40,
307314.
Piepho, H.P., M¨
ohring, J., 2006. Selection in cultivar trials is it ignorable? Crop Sci. 46,
192201.
Piepho, H.P., Laidig, F., Drobek, T., Meyer, U., 2014. Dissecting genetic and non-genetic
sources of long-term yield trend in German ofcial variety trials. Theor. Appl. Genet.
127, 10091018.
Piepho, H.P., Herndl, M., P¨
otsch, E.M., Bahn, M., 2017. Designing an experiment with
quantitative treatment factors to study the effects of climate change. J. Agron. Crop
Sci. 203 (6).
Raman, A., Verulkar, S.B., Mandal, N.P., Variar, M., Shukla, V.D., Dwivedi, J.L., Singh, B.
N., Singh, O.N., Swain, P., Mall, A.K., Robin, S., Chandrababu, R., Jain, A., Ram, T.,
Hittalmani, S., Haefale, S., Piepho, H.P., Kumar, A., 2012. Drought yield index to
select high yielding rice lines under different drought stress severities. Rice 5, 31.
Sadras, V.O., Rebetzkeb, G.J., Edmeades, G.O., 2013. The phenotype and the components
of phenotypic variance of crop traits. Field Crops Res. 154, 255259.
Saikumar, S., Verma, C.M.K., Saiharini, A., Kamleshwer, G.P., Nagendra, K., Lavanya, K.,
Ayyappa, D., 2016. Grain yield responses to varied level of moisture stress at
reproductive stage in an interspecic population derived from Swarna / O.
Glaberrima introgression line. NJAS Wagen. J. Life Sci. 78, 111122.
Saito, Kazuki., 2018. Genetic Gains in Grain Yield of Rice Varieties Released in Sub-
saharan African Countries. International Rice Congress, 2018, Singapore.
Sayre, K.D., Rajaram, S., Fischer, R.A., 1997. Yield potential progress in short bread
wheats in North Mexico. Crop Sci. 37, 3642.
Serraj, R., McNally, K.L., Slamet-Loedin, I., Kohli, A., Haefele, S.M., Atlin, G., Kumar, A.,
2011. Drought Resistance Improvement in Rice: An Integrated Genetic and Resource
Management Strategy. Plant Prod. Sci. 14, 114.
Shaibu, A.A., Urgu, M.I., Sow, M., Maji, A.T., Ndjindjop, M.N., Venuprasad, R., 2017.
Screening African rice (Oryza glaberrima) for tolerance to abiotic stresses: II. Lowland
drought. Crop Sci. 58, 133142.
Sharma, R.C., Crossa, J., Velu, G., Huerta-Espino, J., Vargas, M., Payne, T.S., Singh, R.P.,
2012. Genetic gains for grain yield in CIMMYT spring bread wheat across
international environments. Crop Sci. 52, 15221533.
Singh, S.P., Jain, A., Anantha, M.S., Tripathi, S., Sharma, S., Kumar, S., Prasad, A.,
Sharma, B., Karmakar, B., Bhattarai, R., Das, S.P., Singh, S.K., Shenoy, V., Babu, R.C.,
Robin, S., Swain, P., Dwivedi, J.L., Yadaw, R.B., Mandal, N.P., Ram, T., Mishra, K.K.,
Verulkar, S.B., Aditya, T., Prasad, K., Perraju, P., Krishna, R.M., Sharma, S.,
Anitha, K.R., Kumar, A., Henry, A., 2017. Depth of soil compaction predominantly
affects rice yield reduction by reproductive-stage drought at varietal screening sites
in Bangladesh, India, and Nepal. Pl. Soil 417, 377392.
Smith, A.B., Cullis, B.R., Gilmour, A., 2001. The analysis of crop variety evaluation data
in Australia. Aust. NZ. J. Stat. 43, 129145.
Smith, A.B., Cullis, B.R., Thompson, R., 2005. The analysis of crop cultivar breeding and
evaluation trials: an overview of current mixed model approaches. J. Agric.Sci. 143,
449462.
Tadesse, W., Morgounov, A.I., Braun, H.J., Akin, B., Keser, M., Kaya, Y., Sharma, R.C.,
Rajaram, S., Singh, M., Baum, M., van Ginkel, M., 2013. Breeding progress for yield
in winter wheat genotypes targeted to irrigated environments of the CWANA region.
Euphytica 194, 177185.
Trethowan, R.M., van Ginkel, M., Rajaram, S., 2002. Progress in breeding wheat for yield
and adaptation in global drought affected environments. Crop Sci. 42, 14411446.
Venuprasad, R., Latte, H.N., Atlin, G.N., 2007. Response to direct selection for grain
yield under drought stress in rice. Crop Sci. 47, 285293.
Verulkar, S.B., Mandal, N.P., Dwivedi, J.L., Singh, B.N., Sinha, P.K., Mahato, R.N.,
Dongre, P., Singh, O.N., Bose, L.K., Swain, P., Robin, S., Chandrababu, R., Senthil, S.,
Jain, A., Shashidhar, H.E., Hittalmani, S., Vera-Cruz, S., Paris, T., Raman, A.,
Haefele, S., Serraj, R., Atlin, G., Kumar, A., 2010. Breeding resilient and productive
genotypes adapted to drought-prone rainfed ecosystem of India. Field Crops Res.
117, 197208.
World Bank, 2009. Agriculture and Rural Sector South Asia: Shared Views on
Development and Climate Change. World Bank, Washington, DC, pp. 97107
chapter 7.
Xangsayasane, P., Jongdee, B., Pantuwan, G., Fukai, S., Mitchell, J.H., Inthapanya, P.,
Jothiyangkoon, D., 2013. Genotypic performance under intermittent and terminal
drought screening in rainfed lowland rice. Field Crops Res. 156, 281292.
Xangsayasane, P., Fukai, S., Mitchell, J.H., Jongdee, B., Jothiyangkoon, D.,
Pantuwan, G., Inthapanya, P., 2014. Genotypic performance in multi-location on-
farm trials for evaluation of different on-station screening methods for drought-
prone rainfed lowland rice in Lao PDR. Field Crops Res. 160, 111.
Xu, Y., Li, P., Zou, C., Lu, Y., Xie, C., Zhang, X., Prasanna, B.M., Olsen, M.S., 2017.
Enhancing genetic gain in the era of molecular breeding. J. Exp. Bot. 68, 26412666.
Yamano, T., Mahabayabas, M., Dar, M., 2013. Stress-tolerant Rice in Eastern India:
Development and Distribution. STRASA, Economic Briefs, Bill & Melinda Gates
Foundation, IRRI, Philippines, p. 3.
Zhang, Q.Q., Li, V.P., Singh, P.J., Shi, Q.Z., Sun, P., 2018. Nonparametric integrated
agrometeorological drought monitoring: model development and application.
J. Geophys. Res. 123, 7388.
A. Kumar et al.
... Introgression of prominent QTLs (qDTY 1.1 , qDTY 2.1 , qDTY 2.2 , qDTY 3.1 , qDTY 3.2 , qDTY 6.1 , and qDTY 12.1 ) into high-yielding but droughtsusceptible mega varieties including IR64, MTU1010, TDK1-Sub1, Savitri, Swarna-Sub1, Samba Mahsuri, and Vandana imparted better yield advantage and consistent effects in multiple environments under drought stress in different genetic backgrounds [14][15][16][17][18] . Marker-aided qDTY pyramiding in Indian elite rice varieties (Sahbhagi dhan, DRR Dhan 42, CR Dhan 801, Naveen, and PB 44) showed significantly superior performance under reproductive stage drought stress conditions [19][20][21][22][23] . Improved cultivars with either single or different combinations of qDTYs have already been released in many countries 16 . ...
Article
Full-text available
Haplotype-based breeding is an emerging and innovative concept that enables the development of designer crop varieties by exploiting and exploring superior alleles/haplotypes among target genes to create new traits in breeding programs. In this regard, whole-genome re-sequencing of 399 genotypes (landraces and breeding lines) from the 3000 rice genomes panel (3K-RG) is mined to identify the superior haplotypes for 95 drought-responsive candidate genes. Candidate gene-based association analysis reveals 69 marker-trait associations (MTAs) in 16 genes for single plant yield (SPY) under drought stress. Haplo-pheno analysis of these 16 genes identifies superior haplotypes for seven genes associated with the higher SPY under drought stress. Our study reveals that the performance of lines possessing superior haplotypes is significantly higher (p ≤ 0.05) as measured by single plant yield (SPY), for the OsGSK1-H4, OsDSR2-H3, OsDIL1-H22, OsDREB1C-H3, ASR3-H88, DSM3-H4 and ZFP182-H4 genes as compared to lines without the superior haplotypes. The validation results indicate that a superior haplotype for the DREB transcription factor (OsDREB1C) is present in all the drought-tolerant rice varieties, while it was notably absent in all susceptible varieties. These lines carrying the superior haplotypes can be used as potential donors in haplotype-based breeding to develop high-yielding drought-tolerant rice varieties.
... Research conducted in Eastern India demonstrated that Sahbhagi Dhan has the most consistent yields and outperforms other rice varieties (Dar et al., 2020). Sahbhagi Dhan was created to thrive in drought-prone areas of India and has a yield advantage of 0.5 to 1 ton/ha over other rice varieties during droughts (Kumar et al., 2021). Although it is resistant to leaf blasts, it is only moderately resistant to brown spots, sheath rot, stem borer, and leaf folder. ...
Article
Full-text available
Drought poses a significant challenge to rice cultivation in Asia's rain-fed regions, which is expected to worsen with climate change. This article presents a comprehensive review of the current state of knowledge on drought tolerance in rice, based on a literature review of 52 relevant articles. The articles were chosen based on their relevance to the topic of drought tolerance in rice. The selected articles were then analyzed using a qualitative approach to summarize and synthesize their findings into three main sections: impact, performance, and recent trends. The article highlights several key findings on the development of drought-tolerant rice cultivars, including the identification of genes that control responses to water availability, the use of submergence-tolerant varieties in flood-prone lowlands, and the importance of physiological, biochemical, and molecular adaptation processes in improving rice's tolerance to drought stress. The article emphasizes the importance of marker-assisted breeding and cultivation in semi-arid and rainfed environments to develop more drought-tolerant cultivars. The development of drought-tolerant rice cultivars is crucial to ensure food security and mitigate the effects of climate change in Asia's rain-fed regions. The article also discusses various types of droughts and their effects on different plant species and drought pressures. As the global population increases, the demand for rice as a dependable food crop also rises. To meet this demand, rice cultivation must be expanded to rainfed areas. However, rice's adaptation mechanisms and habitat make it one of the most challenging crops for breeders to develop drought-tolerant varieties. Overall, this article provides important insights and recommendations to improve rice productivity and address the challenges associated with drought in rice cultivation.
Article
Full-text available
The number of panicles per unit area (PNpA) is one of the key factors contributing to the grain yield of rice crops. Accurate PNpA quantification is vital for breeding high-yield rice cultivars. Previous studies were based on proximal sensing with fixed observation platforms or unmanned aerial vehicles (UAVs). The near-canopy images produced in these studies suffer from inefficiency and complex image processing pipelines that require manual image cropping and annotation. This study aims to develop an automated, high-throughput UAV imagery-based approach for field plot segmentation and panicle number quantification, along with a novel classification method for different panicle types, enhancing PNpA quantification at the plot level. RGB images of the rice canopy were efficiently captured at an altitude of 15 m, followed by image stitching and plot boundary recognition via a mask region-based convolutional neural network (Mask R-CNN). The images were then segmented into plot-scale subgraphs, which were categorized into 3 growth stages. The panicle vision transformer (Panicle-ViT), which integrates a multipath vision transformer and replaces the Mask R-CNN backbone, accurately detects panicles. Additionally, the Res2Net50 architecture classified panicle types with 4 angles of 0°, 15°, 45°, and 90°. The results confirm that the performance of Plot-Seg is comparable to that of manual segmentation. Panicle-ViT outperforms the traditional Mask R-CNN across all the datasets, with the average precision at 50% intersection over union (AP50) improved by 3.5% to 20.5%. The PNpA quantification for the full dataset achieved superior performance, with a coefficient of determination (R²) of 0.73 and a root mean square error (RMSE) of 28.3, and the overall panicle classification accuracy reached 94.8%. The proposed approach enhances operational efficiency and automates the process from plot cropping to PNpA prediction, which is promising for accelerating the selection of desired traits in rice breeding.
Article
A sustainable food system ensures food security and nourishment for future generations. Rice (Oryza sativa L.) requires emergence stress tolerance for successful cultivation in germination and early establishment stages. According to the FAO (Food and Agriculture Organization) study, rice is one of the most important worldwide cereal crops and a staple food for more than half of the world’s population. Over the past few years, drastic climatic changes have been subjected to many abiotic and biotic stresses. Global estimates express that a total of 1000 million hectares of cultivated land is affected by abiotic and biotic stress. Besides this around 30% of the irrigated land area is also disturbed. This review discusses the mechanisms of stress-tolerant variants and understanding the recent developments of complex tolerance phenomena. Here, we have compiled the logical pieces as evidence of successful applications of potent molecular tools for boosting stress tolerance in rice, their prospects and limitation. This information would help researchers better understand the genetic enhancement of stress tolerance in rice.
Article
Sustainable rice production in upland habitats depends on achieving higher yields. This study employs correlation and path coefficient analyses to identify essential trait criteria for enhancing rice yield in upland genotypes. The study included two growing seasons using 40 genotypes. Genotypic correlation analysis reveals a robust positive correlation of effective tillering with panicle number and yield. Notably, it shows significant negative correlations with 1000-grain weight and leaf width across diverse locations and cropping seasons. Additionally, the phenotypic estimates underscore a substantial positive correlation between yield and panicle number. Furthermore, the path analysis reveals that panicle number maintains a significantly positive association with yield at the 5% level of significance. Moreover, the analysis of the direct and indirect genotypic effects underscores the significance of culm number, effective tillering, and panicle number, all of which show remarkable and positive correlations with yield, achieving statistical significance at both the 5% and 1% levels. To enhance rice grain yield, a genotype must have an elevated count of pivotal traits per plant, including heightened panicle number, increased panicle length, greater culm number, elongated culm length, a greater number of effective tillers, early flowering initiation, expedited maturation, and augmented leaf length. These characteristics are pivotal determinants contributing significantly to the overall grain yield in rice cultivation and they are instrumental for sustainable rice improvement in the agro-ecology.
Article
Full-text available
Regular measurement of realized genetic gain allows plant breeders to assess and review the effectiveness of their strategies, allocate resources efficiently, and make informed decisions throughout the breeding process. Realized genetic gain estimation requires separating genetic trends from nongenetic trends using the linear mixed model (LMM) on historical multi‐environment trial data. The LMM, accounting for the year effect, experimental designs, and heterogeneous residual variances, estimates best linear unbiased estimators of genotypes and regresses them on their years of origin. An illustrative example of estimating realized genetic gain was provided by analyzing historical data on fresh cassava (Manihot esculenta Crantz) yield in West Africa (https://github.com/Biometrics‐IITA/Estimating‐Realized‐Genetic‐Gain). This approach can serve as a model applicable to other crops and regions. Modernization of breeding programs is necessary to maximize the rate of genetic gain. This can be achieved by adopting genomics to enable faster breeding, accurate selection, and improved traits through genomic selection and gene editing. Tracking operational costs, establishing robust, digitalized data management and analytics systems, and developing effective varietal selection processes based on customer insights are also crucial for success. Capacity building and collaboration of breeding programs and institutions also play a significant role in accelerating genetic gains.
Article
Full-text available
Introduction Addressing the global demand for rice production necessitates innovative approaches to enhance upland rice yield in rainfed agroecosystems, considering the challenges posed by increasing population, limited land fertility, low productivity, and water availability. Methods In this study, our study investigated the impact of biochar and organic fertilizer on ten promising rice lines (G1 – G10) and two control (G11 – G12) cultivars under rainfed conditions. The experimental design used a split-plot design with four soil amendments as main plots, namely control, organic fertilizer, biochar, and biochar + organic fertilizer and 12 rice genotypes as subplot. Results The absolute attainable yield gaps, differentiating organic and control (GAP1), biochar + organic and control (GAP2), and biochar and control (GAP3), ranged from 1.5 to 3.7 or increased of 91–580%, 0.8 to 3.5 (72–560%), and 0.6 to 2.58 tons/ha (58–472%), respectively. Notably, G2 + organic exhibited the highest positive absolute yield gap, ranging from 1.1 to 5.38 tons/ha, based on the yield gap matrix. Furthermore, genotype main effect plus genotype-environment interaction (GGE) biplot analysis identified G2 as the most promising rice line, displaying superior yield performance for cultivation in biochar and organic amended soils. Discussion These findings provide valuable insights for farmers, governments, and stakeholders, offering a roadmap to optimize rainfed areas for rice production, serving as practical guidance to enhance overall rice productivity in rainfed agroecosystems.
Article
Full-text available
Genetic gain has been proposed as a quantifiable key performance indicator that can be used to monitor breeding programs’ effectiveness. The cowpea breeding program at the International Institute of Tropical Agriculture (IITA) has developed and released improved varieties in 70 countries globally. To quantify the genetic changes to grain yield and related traits, we exploited IITA cowpea historical multi‐environment trials (METs) advanced yield trial (AYT) data from 2010 to 2022. The genetic gain assessment targeted short duration (SD), medium duration (MD), and late duration (LD) breeding pipelines. A linear mixed model was used to calculate the best linear unbiased estimates (BLUE). Regressed BLUE of grain yield by year of genotype origin depicted realized genetic gain of 22.75 kg/ha/year (2.65%), 7.91 kg/ha/year (0.85%), and 22.82 kg/ha/year (2.51%) for SD, MD, and LD, respectively. No significant gain was realized in 100‐seed weight (Hsdwt). We predicted, based on 2022 MET data, that recycling the best genotypes at AYT stage would result in grain yield gain of 37.28 kg/ha/year (SD), 28.00 kg/ha/year (MD), and 34.85 kg/ha/year (LD), and Hsdwt gain of 0.48 g/year (SD), 0.68 g/year (MD), and 0.55 g/year (LD). These results demonstrated a positive genetic gain trend for cowpea, indicating that a yield plateau has not yet been reached and that accelerated gain is expected with the recent integration of genomics in the breeding program. Advances in genomics include the development of the reference genome, genotyping platforms, quantitative trait loci mapping of key traits, and active implementation of molecular breeding.
Article
Full-text available
Our objective was to estimate genetic gain for yields in genotypes tested in 11 rice ecosystems from 1995 to 2013 in India and compare the growth trend of human population and national rice grain production in 1974 to 2013. In each ecosystem, the check used remained the same over years but showed similar and significant increases along with top-3 genotypes and experimental mean grain yields derived from sets of genotypes that varied with the year. Therefore, when environmental effects were eliminated, there was no significant genetic gain in yield of genotypes. Annually human population grew linearly at 16.203 million persons and rice production at 1.943 million tonnes (mt) during 1974-1994. This growth slowed during 1995- 2013 in population by 16.131 million persons and in rice production by 1.2753 mt. Breeding for higher genetic yields should be restricted to the four mega environments which offer scope, and exploit the unfolding advancements in rice genomics. The national average yield of un-milled rice was 3.76 t/ha. Evidence indicates that the potential yield in rice is 15-16 t/ha and yields of 10 t/ha is attainable in relatively riskfree irrigated (~20 m ha) and rainfed shallow lowland (11 m ha) ecosystems. Closing yield gap (~6 t/ha) through corrective technological and policy interventions is urgently needed to ensure rice availability to match with the demands of growing population.
Article
Full-text available
For highly productive regions such as Germany, the increase of wheat grain yields observed throughout the 20th century is largely attributed to the progress in crop breeding and agronomic management. However, several studies indicate a strong variability of the genetic contribution across locations that further varies with experimental design and variety selection. It is therefore still unclear to which extent management conditions have promoted the realization of the breeding progress in Germany over the last 100+ years. We established a side-by-side cultivation experiment over two seasons (2014/2015 and 2015/2016) including 16 winter wheat varieties released in Germany between 1895 and 2007. The varieties were grown using 24 different long-term fertilization treatments established in 1904 (Dikopshof, Germany). Averaged over all cultivars and treatments mean yields of 6.88 t ha-1 and 5.15 t ha-1 were estimated in 2015 and 2016, respectively. A linear mixed effects analysis was performed to study the treatment-specific relation between grain yields and year of variety release. Results indicate a linear increase in grain yields ranging from 0.025 to 0.032 t ha-1 year-1 (0.304 to 0.387 % year-1) in plots that were treated with combined synthetic-organic fertilizers without signs of a leveling-off. Yields from low or unfertilized plots do not show a significant progress in yield. Responsiveness of mean yields to fertilizer management increases with year of release and indicates small yield penalties under very low nutrient supply. Results highlight the need to consider the importance of long-term soil fertilization management for the realization of genetic gains and the value of long-term fertilization experiments to study interactions between genetic potential and management.
Article
Full-text available
Flooding is one of the major constraints for rice production in rainfed lowlands, especially in years and areas of high rainfall. Incorporating the Sub1 (Submergence1) gene into high yielding popular varieties has proven to be the most feasible approach to sustain rice production in submergence-prone areas. Introgression of this QTL into popular varieties has resulted in considerable improvement in yield after flooding. However, its impact under non-flooded conditions or years have not been thoroughly evaluated which is important for the farmers to accept and adopt any new version of their popular varieties. The present study was carried out to evaluate the effect of Sub1 on grain yield of rice in different genetic backgrounds, under non-submergence conditions, over years and locations. The study was carried out using head to head trials in farmer's fields, which enable the farmers to more accurately compare the performance of Sub1 varieties with their recurrent parents under own management. The data generated from different head to head trials revealed that the grain yield of Sub1 varieties was either statistically similar or higher than their non-Sub1 counterparts under non-submergence conditions. Thus, Sub1 rice varieties show no instance of yield penalty of the introgressed gene.
Article
Full-text available
High-yielding varieties developed in the 1960s and 1970s at the International Rice Research Institute (IRRI) and elsewhere benefited farmers and the public, ultimately increasing yields and reducing the cost of rice to consumers. Most of these varieties, however, did not have the optimum cooking quality that was possessed by many of the traditional varieties they replaced. In 1985, the IRRI-developed indica variety IR64 was released in the Philippines. In addition to its high yield, early maturity and disease resistance, it had excellent cooking quality, matching that of the best varieties available. These merits resulted in its rapid spread and cultivation on over 10 million ha in the two decades after it was released. It has intermediate amylose content and gelatinization temperature, and good taste. It is resistant to blast and bacterial blight diseases, and to brown planthopper. Because of its success as a variety, it has been used extensively in scientific studies and has been well-characterized genetically. Many valuable genes have been introduced into IR64 through backcross breeding and it has been used in thousands of crosses. Its area of cultivation has declined in the past 10 years, but it has been replaced by a new generation of high-quality varieties that are mostly its progeny or relatives. Continued basic studies on IR64 and related varieties should help in unraveling the complex genetic control of yield and other desirable traits that are prized by rice farmers and consumers.
Article
Full-text available
Abiotic stresses such as droughts and floods significantly constrain rice production in India. New stress tolerant technologies have the potential to reduce yield variability and help insulate farmers from the risks posed by these hazards. Using discrete choice experiments conducted in rural Odisha, we estimate farmers' valuation for drought-tolerant (DT) and submergence-tolerant (SubT) traits embodied in rice cultivars. Our results demonstrate that farmers in both drought-prone as well as submergence prone regions value reduction in yield variability offered by new, stress-tolerant cultivars, and would generally be willing to pay a significant premium for these traits. While virtually all farmers perceive the threat of drought and are willing to pay for protection against drought risk, only farmers in flood-prone areas would be willing to pay for rice that can withstand being submerged for prolonged periods, suggesting the potential for market segmentation along geographical or ecological lines.
Article
Full-text available
Drought is a major rice production constraint in Sub Saharan Africa region. Oryza glaberrima, the cultivated rice species which originated from West Africa is well-adapted to its growing ecologies. This study was initiated to identify promising O. glaberrima accessions tolerant to lowland drought stress from the 2106 accessions held at the AfricaRice gene bank. Screening was done over a three-year period in West Africa using standardized protocol and involved evaluating for grain yield under drought and/or irrigated conditions, selecting the high-yielding lines and repeating the testing with the newly selected lines. Four accessions (TOG 7400, TOG 6520, TOG 6519-A and TOG 7442-B) with consistently higher grain yield under drought stress and irrigated conditions were selected. These four accessions originated from three countries in West Africa, namely, Ghana, Liberia and Nigeria. The selected O. glaberrima accessions could be used as donors in breeding for drought tolerance in rice.
Article
Full-text available
Genetic gain within the CIMMYT Eastern and Southern Africa (ESA) hybrid maize (Zea mays L.) breeding program from 2000 to 2010 was recently estimated at 0.85 to 2.2% yr⁻¹ under various environmental conditions. Over 100 varieties were disseminated from CIMMYT to farmers in ESA, hence the need to check genetic diversity and frequency of use of parents to avoid potential narrowing down of the genetic base. Fifty-five parents from CIMMYT ESA used in the hybrids were fingerprinted using genotyping-by-sequencing. Data analysis in TASSEL and MEGA6 generated pairwise genetic distances between parents of 0.004 to 0.4005. Unweighted pair group method with arithmetic mean (UPGMA) analysis produced two clusters (I and II) with two subclusters each (A and B) and two sub-subclusters (IAi and IAii). Principal coordinate analysis produced three clusters where IAi and IIA from the UPGMA analysis formed independent clusters while IAii, IB, and IIB clustered together. Lines were separated by pedigree and origin. Ninety-five percent frequency of pairwise genetic distances ranged between 0.2001 and 0.4000. However, only four of the 55 parents (CML444, CML395, CML312, and CML442) were each used in 15 to 30 of the 52 hybrids evaluated in the genetic gain study. The remaining 51 were used in one to four hybrids. Frequent use of the four parents gave 29 to 58% of the hybrids a narrow genetic base, posing risk in case of pest or disease outbreaks. Parents evaluated do not represent the genetic base of CIMMYT ESA but parents of the best-performing hybrids selected from 2000 to 2010. Breeders should ensure a wide genetic base for released varieties to avoid breakdown in case of pest or disease outbreaks.
Article
Full-text available
Experiments for studying the effects of climatic change on ecosystems often involve manipulation of one or several quantitative treatment factors of interest. Response surface regression is the method of choice for these types of experiment. Here, we describe the development of a design of a free air CO2 enrichment experiment with two quantitative treatment factors, that is, elevated temperature and CO2 enrichment. The design strategy takes account of budget constraints imposing limitations on the number of plots with elevated temperature and CO2 levels. The approach is based on polynomial regression models and is focussed on an efficient estimation of interaction between the two treatment factors. Extension to more than two factors is straightforward. An analysis of soil moisture data demonstrates the overall suitability of the proposed design to analyse non-linear interactions of two (or more) global change factors.
Article
Metric learning has become a widespreadly used tool in machine learning. To reduce expensive costs brought in by increasing dimensionality, low-rank metric learning arises as it can be more economical in storage and computation. However, existing low-rank metric learning algorithms usually adopt nonconvex objectives, and are hence sensitive to the choice of a heuristic low-rank basis. In this paper, we propose a novel low-rank metric learning algorithm to yield bilinear similarity functions. This algorithm scales linearly with input dimensionality in both space and time, therefore applicable to high-dimensional data domains. A convex objective free of heuristics is formulated by leveraging trace norm regularization to promote low-rankness. Crucially, we prove that all globally optimal metric solutions must retain a certain low-rank structure, which enables our algorithm to decompose the high-dimensional learning task into two steps: an SVD-based projection and a metric learning problem with reduced dimensionality. The latter step can be tackled efficiently through employing a linearized Alternating Direction Method of Multipliers. The efficacy of the proposed algorithm is demonstrated through experiments performed on four benchmark datasets with tens of thousands of dimensions.
Article
Drought is a major natural hazard that has massive impacts on the society. How to monitor drought is critical for its mitigation and early warning. This study proposed a modified version of the multivariate standardized drought index (MSDI) based on precipitation, evapotranspiration, and soil moisture, i.e. modified multivariate standardized drought index (MMSDI). This study also used nonparametric joint probability distribution analysis. Comparisons were done between standardized precipitation evapotranspiration index (SPEI), standardized soil moisture index (SSMI), MSDI, and MMSDI, and real-world observed drought regimes. Results indicated that MMSDI detected droughts that SPEI and/or SSMI failed to do. Also, MMSDI detected almost all droughts that were identified by SPEI and SSMI. Further, droughts detected by MMSDI were similar to real-world observed droughts in terms of drought intensity and drought-affected area. When compared to MMSDI, MSDI has the potential to overestimate drought intensity and drought-affected area across China, which should be attributed to exclusion of the evapotranspiration components from estimation of drought intensity. Therefore, MMSDI is proposed for drought monitoring that can detect agrometeorological droughts. Results of this study provide a framework for integrated drought monitoring in other regions of the world and can help to develop drought mitigation.