-
Agriculture, Ecosystems and Environment 233 (2016) 325–335
Conservation agriculture-based wheat production better copes
withextreme climate events than conventional tillage-based systems:
Acase of untimely excess rainfall in Haryana, India
Jeetendra Prakash Aryal, Climate Economista,*,Tek Bahadur
Sapkota, Mitigation Agronomista, Clare Maeve Stirling, Senior
Agronomistb,M.L. Jat, Senior Cropping Systems Agronomista, Hanuman
S. Jat, Senior Agronomistc,Munmun Rai, Senior Agronomista, Surabhi
Mittal, Senior Agricultural Economista,Jhabar Mal Sutaliya, Senior
Agronomistc
a International Maize and Wheat Improvement Center (CIMMYT), CG
Block, National Agricultural Science Center (NASC) Complex, DPS
Marg, Pusa Campus,New Delhi 110012, Indiab International Maize and
Wheat Improvement Center (CIMMYT), Texcoco, Mexicoc International
Maize and Wheat Improvement Center (CIMMYT), CIMMYT, Karnal,
Haryana, India
A R T I C L E I N F O
Article history:Received 6 February 2016Received in revised form
4 September 2016Accepted 13 September 2016Available online xxx
Keywords:Conservation agriculture-based wheatproduction
systemConventional tillage-based wheatproduction systemClimatic
extremesRainfall variabilityIndia
A B S T R A C T
This study explores whether conservation agriculture-based wheat
production system (CAW) can bettercope with climatic extremes than
the conventional tillage-based wheat production system (CTW).
Toassess this, we used data collected from 208 wheat farmers in
Haryana, India in 2013–14 (a period withnormal rainfall i.e.,
normal year) and 2014–15 (a period with untimely excess rainfall
i.e., bad year) wheatseasons. Our analysis shows that whilst
average wheat yield was greater under CAW than CTW duringboth bad
and normal years, the difference was two-fold greater during the
bad year (16% vs. 8%). Thisprovides new evidence that CAW can cope
better with the climatic extremes, in this case untimely
excessrainfall, compared to CTW. Absolute yield of the CAW and CTW
was 10% and 16% lower in the bad yearcompared to the normal year,
respectively. Extreme climate events, such as excess rainfall
during wheatseason, can occur once in every four years in Haryana
and result in a loss of income to both farmers,through a loss of
yield, and the government, through compensatory payments to
farmers. If, as targetedby the Haryana government in 2011, one
million ha of wheat was brought under CAW, the state wouldhave
produced an additional 0.66 million Mg of wheat in 2014–15,
equivalent to US$ 153 million. This isan important finding given
the increased vulnerability of wheat production to climatic
variability in thisregion.
ã 2016 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Agriculture, Ecosystems and Environment
journal homepage: www.elsev ier .com/locate /agee
1. Introduction
Wheat plays a dominant role in global food security as
itcontributes almost 20% of the total dietary calories and
proteinsworldwide and almost 24% in South Asia (Shiferaw et al.,
2013). InIndia, wheat is grown on about 29 million ha and is an
importantcrop for food security. As India accounts for 12% of
global wheatproduction, any loss of production in the region will
have majorrepercussions for global food security (FAO, 2013).
Wheatproductivity and total production in India increased
tremendously
* Corresponding author.E-mail addresses: [email protected],
[email protected] (J.P. Aryal).
http://dx.doi.org/10.1016/j.agee.2016.09.0130167-8809/ã 2016
Elsevier B.V. All rights reserved.
with the advent of green revolution (GR). However maintaining
thegains of GR is increasingly a challenge with wheat yield in
Indiahaving plateaued for last couple of years. Many factors such
asdeclining soil fertility, degrading natural resources and
increasingcost of production inputs are responsible for the recent
stagnationof wheat yield, further compounded by the effects of
climatechange and climatic variability.
Climatic variability in terms of rainfall (drought, excess
rains)and terminal heat (i.e. high temperature during grain filling
stage)severely impact on wheat production in India. For example,
ifwheat is planted late, high temperature at the end of
seasonhastens maturity and reduces grain yield (commonly known
asterminal heat stress). One solution is to plant the wheat crop
early
http://crossmark.crossref.org/dialog/?doi=10.1016/j.agee.2016.09.013&domain=pdfmailto:[email protected]:[email protected]://dx.doi.org/10.1016/j.agee.2016.09.013http://dx.doi.org/10.1016/j.agee.2016.09.013http://www.sciencedirect.com/science/journal/01678809www.elsevier.com/locate/agee
-
Fig. 1. Distribution of rainfall (weekly total) over the wheat
growing season for the years 2013–14 and 2014–15 and long-term
average of weekly total (1982–2013). The errorbar in the long-term
average shows the standard deviation. DAS = Days after sowing.
326 J.P. Aryal et al. / Agriculture, Ecosystems and Environment
233 (2016) 325–335
in November to escape terminal heat but this increases the risk
ofexposure to heavy rainfall during late February and March
(whichis largely unpredictable). Therefore, farmers need to adjust
thetime of sowing as well as the wheat production system.
Adverse impacts of climatic variability on crop productivity
hasbecome increasingly common in India. For instance, in 2004, due
tohigh temperature wheat matured 10–20 days earlier than
normalleading to a loss of more than 4 million Mg of wheat
production(Samra and Singh, 2004). More recently in 2009–10, an
abruptincrease in temperature during the grain filling stage of
wheat wasassociated with an average yield loss of about 6% in
north-eastIndia (Gupta et al., 2010). To ensure food security there
is a need tofocus on resource-efficient technologies that maximizes
crop yieldand increase adaptation to climatic variability.
Conservation agriculture-based wheat production system(CAW) is
based on the principle of minimum soil disturbance,permanent soil
cover and crop diversification (intercropping and/or rotation). In
the rice-wheat system, CAW offers a means ofadvancing sowing date
of wheat compared with the conventionaltillage-based wheat
production system (CTW) due to the timesaved with direct seeding.
The CAW was introduced in India in1990s as one of the
resource-conserving technologies under theRice-Wheat Consortium
(Harrington and Erenstein, 2005). Despitea plethora of studies to
compare CAW with CTW in terms of soilproperties, productivity,
resource use efficiency, economic profit-ability and environmental
sustainability (Aryal et al., 2015c;Erenstein and Laxmi, 2008;
Erenstein et al., 2008; Gathala et al.,2011; Gupta et al., 2010;
Jat et al., 2014; Khatri-Chhetri et al., 2016;Sapkota et al.,
2015), there is still a dearth of documented evidenceof the
adaptive and risk-bearing capacity of CAW under
climaticextremes.
This study assesses whether CAW better copes with
untimelyrainfall compared with CTW and whether the yield benefits
of CAWvary with farm size. We compare the yield under CAW and CTW
ina bad year to assess the adaptive capacity of CAW. In this
study,CAW refers to the zero tillage (ZT) based wheat production
systemwhere residues from the previous crop (mainly rice residue in
thiscase) are retained on the soil surface. We did not consider ZT
in theabsence of residue retention. We selected the state of
Haryana,
India as the study area for the following reasons. Firstly, CAW
hasbeen practiced in Haryana for more than two decades andtherefore
provided an appropriate region for a comparisonbetween CTW and CAW.
Secondly, in the year 2014–15, wheatproduction in Haryana suffered
severely due to untimely andexcessive rainfall (Global Watch and
FAO, 2015). As a result, theGovernment of Haryana provided Indian
rupees11092 crore (i.e. US$ 174.72 million) as compensation to
farmers in 2015 (TheEconomic Times, 2015). Thirdly, in our informal
discussions withfarmers in Haryana, most of them claimed that CAW
suffered lessfrom untimely rainfall compared with CTW. All these
factorscombined to provide a suitable set of conditions to test
theadaptive response of CAW under ‘real life’ climatic
extremes.
2. Effects of excess rainfall on wheat production
More than 80% of the total area under wheat production in
Indiais irrigated (Kumar et al., 2004). Irrigation schedule for
wheat ismore or less standardized. Farmers in western IGP
generallyirrigate wheat 3–4 times depending on the management
practicesand soil type. Usually the irrigations are timed at the
crown rootinitiation stage (20–25 days after sowing, DAS), active
tilleringstage (40–45 DAS), the flowering stage (70–75 DAS) and
grainfilling stage (110–120 DAS). Winter rainfall during
vegetativegrowth of wheat is generally beneficial but too much
during thegrain filling to physiological maturity can be harmful.
Farmersirrigate their crop on a routine basis but winter rains are
largelyunpredictable and rainfall immediately after irrigation is
the mostdetrimental as it results in prolonged water logging which
resultsin a yellowing of leaves and stunted growth. The problem can
becompounded by farmers over fertilizing the water-damaged
crop,believing the yellowish-green crop stand is due to N
deficiency.Excessive N application increases the susceptibility to
diseases,pests and crop lodging. Water stagnation during
grain-filling stagealso causes blackening of the wheat ear-head and
loss of grain fill. Ifgrain filling occurs, grains are shriveled
and light weighted, whichleads to substantial yield loss. Rainfall
during maturity delays
1 The exchange rate for the year 2015 is: Indian rupees (INR) 1
= US$ 0.016
-
J.P. Aryal et al. / Agriculture, Ecosystems and Environment 233
(2016) 325–335 327
harvesting of wheat as farmers have to wait until the
soilconditions and crop stand are suitable for operating the
combineharvester. This will not only affect the wheat output but
also affectsthe wheat procurement process. Rains during the
ready-to-harvestperiod can increase the grain moisture content to
above 14%making it unsuitable to sell directly from the field as
per theexisting norms of Food Corporation of India.
In conventional systems, soils are generally puddled to
restrictdrainage during the rice season but this creates a hard pan
whichfurther restricts vertical water movement during the wheat
seasoncreating conditions prone to waterlogging. By contrast, the
ZTsystem in which residues from the previous crop are retained in
thesoil not only help conserve moisture during drought conditions
butalso enhance infiltration and percolation of water in the event
ofexcess or untimely rainfall.
In the conventional system of production, wheat seeds
arebroadcasted followed by roto-tillage and planking. This results
insome seeds remaining on the surface while others are buried
deepinto the soil. Whilst most seeds germinate well, those nearest
thesurface may not develop a sufficiently deep rooted system that
isable to withstand the impact of heavy rain and wind and so
areprone to lodging. In ZT, on the other hand, seeds and fertilizer
aredrilled at a consistent and optimal depth and
row-geometryresulting into a well-developed root system (Singh et
al., 2014) thatconfers greater resilience against adverse rainfall
conditions.
3. Untimely rainfall and wheat crop loss in Haryana
In general the rice-wheat (RW) growing area of Haryana has
asemi-arid subtropical climate, characterized by very hot
summersand cool winters. Historical data indicates that there
exists hugevariability in total annual rainfall ranging from 350 to
1400 mmwith average annual rainfall of about 750 mm, 75% of which
isreceived during June–September based on long-term data
fromCentral Soil Salinity Research Institute (CSSRI) Karnal,
Haryana. Asthe wheat crop is grown during winter season
(November–April),variability of monsoon rainfall has little effect
on growth and yieldexcept at the end of the monsoon when rainfall
affects wheatseeding. By contrast, variability in winter rainfall
strongly affectsgrowth and production of wheat. Analysis of
long-term weather
Fig. 2. Maximum, minimum and average weekly temperature during
the wheat growingdotted vertical lines represent the grain
development stage which is sensitive to high
data (1982–2013) from the meteorological department of
CSSRIKarnal reveals that the total amount of rainfall received
during thewheat season (29 November to 15 April) varies from 10 to
300 mm(average 110 mm). Not only is the amount of seasonal
rainfallimportant but also the distribution in relation to crop
growth withwheat performing better if rainfall is uniformly
distributed overthe growing period rather than received in a few
torrentialdownpours. In this respect, the 2013–14 and 2014–15
wheatseasons were considered normal and abnormal respectively
basedon the seasonal rainfall distribution.
Total rainfall received during wheat season (29 October to
15April) in 2013–14 and 2014–15 was 170 and 235 mm,
respectively(Fig. 1). Whilst rainfall in the 2013–14 wheat season
was uniformlydistributed between the late vegetative to early
maturity stage, inthe 2014–15 season 68% (160 out of 235 mm) of the
total rainfallwas received only in two windows of one week duration
each i.e.113–120 DAS (critical stage of grain development) and
149–155DAS (ready to harvest stage). In both years, terminal heat
stress wasnot observed as temperature was below the threshold level
of 33 �Cduring the grain filling stage (Fig. 2). Minimum and
maximumtemperature for the wheat growing period was more favorable
in2014–15 than in 2013–14 and yet the former was a bad year
forwheat production in Haryana and this can be attributed more
thananything to untimely and excess rainfall. Analysis of
long-termweather data (1982–2015) from the study area reveals that
badweather in terms of excess rainfall during critical stages of
wheatoccurs one in every four years. Here, any year that received
morethan 100 mm rainfall between 54 and 105 day of the year
(graindevelopment and maturity period of wheat) was considered
badyear.
Government of Haryana reported wheat yield loss of 5.8%
in2014–15 compared with 2013–14 (Table 1) with the highest
lossrecorded in the Karnal district (17.6% loss over 2013–14)
duemainly to incessant rainfall over a few days during
graindevelopment and ready-to-harvest stage.
4. Study area and data
For this study, we used the data collected from 208 wheatfarmers
in 10 village clusters (scattered over 15 villages) in the
seasons in 2013–14 and 2014–15. In each growing season the
window between twotemperature. DAS = Days after sowing.
-
Table 1Area, production, yield and percentage yield loss in
wheat during 2013–14 and 2014–15.
State and district Area (million ha) Production (million Mg)
Yield (Mg ha�1) Yield loss (%) over 2013–14
2013–14 2014–15 2013–14 2014–15 2013–14 2014–15
Haryana 2.499 2.54 11.8 11.3 4.722 4.45 5.8Karnal 0.172 0.174
0.845 0.704 4.912 4.046 17.6
Source: Department of Agriculture, Government of Haryana
(http://www.agriharyana.nic.in accessed on 20 November 2015).
328 J.P. Aryal et al. / Agriculture, Ecosystems and Environment
233 (2016) 325–335
Karnal district, Haryana for two consecutive wheat seasons
of2013–14 and 2014–15 (Fig. 3; Table 2).
Of the total number of farm households selected for the
study,half of the households are selected from those who have
adoptedCAW (i.e., adopters of CAW) and the remaining half was
selectedfrom those who have not adopted it at all (i.e.,
non-adopters ofCAW). The data comprises of information on major
householdcharacteristics, total operated land, land areas under CAW
andCTW, production inputs, crop management and grain yield underCAW
and CTW.
4.1. Descriptive statistics of the study households
Table 3 presents the characteristics of the sample
households.Average age of household heads is about 40 years for
both adoptersand non-adopters of CAW. All of the sample households
reportedthat farming is their primary occupation. Very few farm
house-holds have secondary occupations which include dairy,
business,and rice mills. Average land holding size for adopters is
6.6 hacompared with only 4.8 ha for non-adopters of CAW. The
majorityof the sample households have their own tractors. Among
CAWadopted households, only five have their own ZT machine
capableof seeding over previous crop residues (i.e. Turbo Happy
Seeder).Therefore, most of the farmers rely on the custom hiring
service forthe machines required for CAW. Of the total adopters,
approxi-mately 30% of households have taken some kind of training
inconservation agriculture.
5. Empirical framework of the study
The empirical framework of the study is as follows:
5.1. Analysis of major factors affecting wheat yield under
normal andbad years
To test whether CAW better copes with extreme rainfallcompared
with CTW, we classified all sample farm households intotwo
categories (i.e., adopters of CAW and non-adopters of CAW)based on
the type of wheat production system followed. Weestimated multiple
regression model with dummy variable so thatwe can control for the
impact of other inputs/farm managementelements that may affect
wheat yield. The multiple regressionmodel with wheat production
system dummy can be presented as:
y ¼ xb þ b1D1 þ e; e!N ð0; 1Þ ð1Þ
In Eq. (1), y is the wheat yield (Mg per ha), x is the vector of
allother explanatory variables other than wheat production
system(i.e., seed variety, number of irrigations applied, sowing
dates,fertilizers applied, education of the household head,
participationin agricultural training and access to credit for
agriculturalactivities) and D1 refers to dummy variable which takes
value 1if farmer has produced wheat using CAW and 0 otherwise. b
and b1are coefficient vector and coefficient to dummy variable,
respec-tively while eis the stochastic error term.
This analysis provides us with a basis to isolate the
yielddifference due to other factors and hence, helps to
attribute
whether the yield difference is due to differences in the
wheatproduction systems. In addition, it also tested whether yield
underCAW is higher than yield under CTW during a bad year,
therebytesting the adaptive or climate risk coping capacity of
CAW.
5.2. To test the yield difference between CA-based and
conventionaltillage-based wheat production systems
We applied t-test of significance difference between twosample
means to check whether there is difference between thesetwo
alternative production systems. This test is carried out for
boththe normal (2013–14) and bad year (2014–15).
We also used stochastic dominance analysis (SDA) to comparethe
wheat yield distribution between CAW and CTW. In SDA, thecumulative
distribution functions (CDFs) of yield of the wheat cropunder
alternative systems are compared for the normal and badyear
separately. Two major criteria for comparing the
stochasticdominances are: first-order stochastic dominance (FSD)
andsecond-order stochastic dominance (SSD). Assume that CAW(y)and
CTW(y) are cumulative distribution functions of wheat yieldsfor
CA-based and conventional tillage-based wheat productionsystems
respectively. Under the FSD criterion, the distribution CAW(y)
dominates CTW(y) if CTWðyÞ � CAWðyÞ � 0; 8y �
-
Fig. 3. Study area.
J.P. Aryal et al. / Agriculture, Ecosystems and Environment 233
(2016) 325–335 329
-
Table 2Villages under study and the distribution of sample
size.
Villages Sample size
Anjanthali/Balu 28Badarpur/Dabkolikala 24Bastada/Kutail
14Chorpura 20Daha/Uncha Saman 18Gangar 20Padhana/Sandhir 24Pujam
20Shambli 20Taraori 20
330 J.P. Aryal et al. / Agriculture, Ecosystems and Environment
233 (2016) 325–335
For assessing the robustness of the results, we also
estimatednon-parametric regressions to look at the association
betweenyields and farm size under the two different wheat
productionsystems in both normal and bad years. In general,
non-parametricregression methods fit a local relationship between
the dependentvariable y and the regressor x. The local relationship
refers to theseparate fitted relationships that are obtained at
different values ofx (Cameron and Trivedi, 2009).
Consider a local linear regression model:y ¼ m xð Þ þ u, where
m(.) is the conditional mean function and x is a scalar. A
localregression estimate of m(x) at x = x0 is a local weighted
average of yi,i = 1, 2, . . . .,N, that places greater weight on
observations while xiis closer to x0 and less weight on
observations while xi is far from x0.This can be represented
by:
m̂ x0ð Þ ¼XN
i¼1 wðxi; x0; hÞyiwhere w(xi, x0, h) represents the weight which
decreases when thedistance between xi and x0 increases. As the
bandwidth parameterh increases, more weight is placed on
observations for which xi isclose to x0. The local linear estimator
additionally includes a slopecoefficient and at x = x0
minimizes,XN
i¼1 Kxi � x0
h
� �yi � a0 � b0 xi � x0ð Þ
� �2
where K(.) is a kernel function that places greater weights on
pointsxi is close to x0. The local linear estimator with degree
ofpolynomial (t) greater or equal to 1 does much better than
thepreceding methods at estimating m(x0) at values of x0 near
theendpoints of the range of x, as it allows for any trends near
the endpoints. Therefore, we used t ¼ 1 in our estimation. Of the
several
Table 3Characteristics of sample households.
Household (HH) characteristics
Age of HH head (yr.) Illiterate HH head (no.) HH head with up to
secondary education (no.) HH head with above secondary education
(no.) Average land size (in ha) HHs with own tractor (no.) HHs with
own laser land leveler (no.) HHs with own ZT machine (no.) HHs
participated in CA training (no.) HHs participated in soil and
water management training Access to credit required for agriculture
Membership in farmer cooperatives Membership in any other
institutions Know about climate change Have secondary occupation
Total sample size
variants of nonparametric regressions, we estimated a
localpolynomial regression, a variation of local regression,
mainlybecause it better explains the variations in data (for
details, seeCameron and Trivedi, 2009). We also checked kernel
densities ofyield functions for these two alternative production
systems andtested for the equality of the two distributions using
two-sampleKolmogorov-Smirnov tests.
6. Results
6.1. Factors determining yield difference in normal and bad
year
CAW had a positive and statistically highly significant effect
onwheat yield compared to CTW in both bad and normal years(Table
4). In the normal year only one other variable – amount ofurea
applied – was found to have a significant and positive impacton
wheat yield, whereas in the bad year DAP and urea had anegative
impact on yield. This is probably because, pale yellowishand
stunted growth of plant due to water stagnation (in bad year)may
have given an impression of under-fertilization and as a
resultfarmers might have applied more fertilizer. We tested
whetheramount of fertilizer applied was different between CAW and
CTWin normal and bad year separately. We found no
significantdifference in the application of fertilizer in both
cases. This furtherhelps us to attribute CAW to observed yield
differences.
Of the socio-economic variables, better educated farmers
arefound to have slightly higher yields compared to illiterate
farmersin both normal and bad years. Similarly, yields were higher
forfarmers who had received agricultural training compared
withthose that had not. Whilst there is always the option to add
morevariables, the results presented in Table 4 justify the
classificationof sample households in terms of adopters and
non-adopters ofCAW.
6.2. Yield difference between CAW and CTW
Wheat yield was 8.1% higher in the CAW than CTW in thenormal
year and this difference was found to be statisticallysignificant
at 99% confidence level (Table 5). Of significance is
theobservation that the yield advantages of CAW over CTW was
muchhigher (15.6%) (and again statistically significant) in the
2014–15bad year. As far as we are aware, these results are the
first of theirkind, providing clear evidence of the benefits of CAW
in conferringresilience to climatic extremes. Another important
finding is thatthe yield loss in the bad year (compared to normal
year) was less inCAW (10.4%) compared with CTW (16.2%) indicating a
greater yield
Adopters of CAW Non-adopters of CAW
39.4 40.59 1344 6451 276.6 4.885 715 014 031 25 0104 8637 11
0104 1045 0104 104
-
Table 4Factors explaining yield in normal and bad years.
Explanatory variables Normal year Bad year
Variety dummy (1 if H2967, 0 otherwise) �0.029 0.115(0.090)
(0.087)
Production system dummy (1 if CAW, 0 if CTW) 0.392***
0.617***(0.088) (0.082)
DAP (kg per ha) 0.007 �0.014***(0.007) (0.005)
Urea (kg per ha) 0.002** �0.003***(0.001) (0.001)
Number of irrigation applied �0.046 0.060(0.074) (0.121)
Potash (kg per ha) 0.005 0.002(0.003) (0.003)
Sulphur (kg per ha) �0.016 �0.008(0.010) (0.010)
Date of sowing �0.001 �0.004(0.006) (0.008)
HH head with up to secondary education (base category:
illiterate) 0.094** 0.101***(0.045) (0.037)
HH head with above secondary education (base category:
illiterate) 0.048*** 0.083***(0.014) (0.031)
Access to credit dummy (Yes = 1, No = 0) �0.129 0.114*(0.110)
(0.067)
Participated in agricultural training (Yes = 1, No = 0) 0.043***
0.065***(0.017) (0.023)
Constant 4.179** 3.993***(2.002) (1.369)
R-Squared 0.32 0.51No. of Observation 208 208
Note: *, **, and *** refer to 10%, 5%, and 1% level of
significance, respectively. Standard errors are reported in
parentheses.
J.P. Aryal et al. / Agriculture, Ecosystems and Environment 233
(2016) 325–335 331
penalty of the latter under climatic extremes. This was
confirmedby the stochastic dominance analysis (Fig. 4) where it was
possibleto use the results of the first-order analyses because the
CDFs didnot intersect one another. In both years, CDFs for CAW were
belowthat for CTW, indicating that the former dominated the
later.
Table 5Wheat yield (Mg/ha) under alternative production systems
in normal and bad year.
Wheat production system Average wheat yield (Mg/ha)
2013–14(Normal year)
2014–15(Bad year
CAW 5.46 4.89 (0.0517) (0.0558)
CTW 5.05 4.23 (0.0583) (0.0662)
Yield difference between CAW and CTW (Mg/ha) 0.41c 0.66d
t-test 5.19*** 7.59***
a yield loss due to climatic risk (untimely rain at the ready to
harvest stage of wheab Yield loss due to climatic risk (untimely
rain at the ready to harvest stage of wheac Yield gap between CAW
and CTW in normal year i.e., in year 2013–14.d Yield gap between
CAW and CTW in bad year (here, year with untimely rain at the re***
Refers to significant at 99% confidence level; observations are
combined observat
However, the difference between CAW(y) and CTW(y) was larger
inthe bad year, implying that CAW had a greater yield potential
orlesser yield penalty in the bad year compared to CTW.
Yield difference between normal year and bad year (Mg/ha)
t-test
)
0.57a 7.55***
0.82b 9.38***
t crop) in CAW between year 2013–14 and year 2014–15.t crop) in
CTW between year 2013–14 and year 2014–15.
ady to harvest stage of wheat crop) i.e., year 2014–15 – address
climatic variability.ions; standard errors are reported in
parentheses.
-
Fig. 4. Stochastic dominance analysis of the wheat yield
difference between CTW and CAW in a normal and bad year.
332 J.P. Aryal et al. / Agriculture, Ecosystems and Environment
233 (2016) 325–335
6.3. Yield variation across different farm sizes
Yield differences between CAW and CTW were higher
andstatistically significant across all farm size categories in the
badyear, whereas this difference was not statistically significant
formarginal and small farmers in the normal year (Table 6).
The results of non-parametric regression (local
polynomialregression analysis) is shown in Fig. 5. As compared to
CAW, yieldvariation was higher in CTW in both normal and bad years
and wasmuch higher among larger farms in the bad year (Fig. 5,
rightpanel). This means shifting from CTW to CAW is more crucial
forlarge farmers in the region and given their contribution
tomarketable surplus, this adaptive response also has
importantimplications for food security.
We also carried out the kernel density functions for the
yieldfunctions under the CAW and CTW and used the Kolmogorov-
Table 6Yield variation across farm size in normal and bad year
under CAW and CTW.
Land size (in ha) Yield (Normal year): 2013–14
CAW CTW Diff.
Marginal (�1) 5.63 5.25 0.38 (0.125) (0.211)
Small (>1 and �2) 5.62 5.11 0.52 (0.217) (0.135)
Semi-medium (>2 and �4) 5.53 5.193 0.34 (0.094) (0.081)
Medium (>4 and �10) 5.36 4.97 0.39 (0.078) (0.062)
Large (>10) 5.59 5.27 0.33 (0.103) (0.096)
Note: ** and *** refer to significant at 95% and 99% confidence
levels respectively. Standaron the FAO (for details, visit
http://www.fao.org/ag/agp/agpc/doc/counprof/India/India.
Smirnov test to confirm that differences in yield distributions
werestatistically significant.
6.4. Farmers’ perception on why CAW performs better than CTW
undervariable and untimely rainfall conditions
In response to the question whether or not you would like
tocontinue with CAW, all of the adopters responded positively and
allbelieved that this system copes better with variable and
untimelyrainfall during the wheat season.
All farmers, who adopted CAW believe that a better root systemis
the major reason why it copes better with untimely rainfall
andalmost 35% of adopters considered that better water infiltration
inthe CAW compared with CTW system reduces yield losses (Table
7).
t-test Yield (Bad year): 2014–15 t-test
CAW CTW Diff.
0.896 5.38 4.69 0.68 3.29***(0.125) (0.102)
1.53 5.07 4.19 0.87 2.68***(0.157) (0.143)
2.46** 5 4.26 0.75 5.22***(0.093) (0.085)
3.91*** 4.77 4.11 0.65 6.41***(0.087) (0.058)
2.29** 5 4.31 0.69 4.13***(0.099) (0.119)
d errors are reported in the parentheses. We classified the
sample households basedhtm).
http://www.fao.org/ag/agp/agpc/doc/counprof/India/India.htm
-
Fig. 5. Yield variations across different farm sizes under two
alternative wheat production systems.
Table 7Reasons why CAW copes with untimely rainfall.
Reasons why CAW copes better Adopters
Better water percolations 36Better root system 104Better
fertilizer management 3Total sample size 104
J.P. Aryal et al. / Agriculture, Ecosystems and Environment 233
(2016) 325–335 333
7. Constraints to adoption of CAW
Table 8 presents the major constraints to adopt CAW reportedby
the sample households.
Lack of machine availability is the major constraint to
farmeradoption and this problem is also related with the type of
ZTmachine that is required for sowing wheat with rice residue in
thefield (Sidhu et al., 2007). In addition, 68% of non-adopters
reportedthat lack of the knowledge is a major constraint to the
uptake ofCAW by farmers, similar to the findings of Sapkota et al.
(2015).
Table 8Major constraints to adopt CAW.
Constraints Adopters Non-adopters
Lack of machine availability 67 103Lack of knowledge 9 71Lack of
credit facility 0 18Lack of confidence on CAW 1 13Total sample size
104 104
8. Discussions and policy implications
This study provides new and important evidence of the benefitsof
conservation agriculture-based wheat production system (CAW)in
terms of resilience to untimely rainfall during the wheat
seasoncompared to conventional tillage-based wheat production
system(CTW). Furthermore, CAW performs better under both normal
andbad years. As farmers in many states in India face similar
andincreased risks of climatic extremes such as untimely
heavyrainfall, these findings have important implications for
designingpolicies to improve adaptation of agriculture to climate
change.The findings are discussed from the farmer’s and
government’sperspectives.
8.1. Implications for farmers
Compared to CTW, the additional wheat yield obtained underCAW is
0.41 Mg ha�1 in a normal year and 0.66 Mg ha�1 in a badyear. This
means a farmer can benefit from an additional amount ofUS$ 95 ha�1
(i.e., 0.41 Mg ha�1� US$ 232 Mg�1) in a normal yearand US$ 153 ha�1
(i.e., 0.66 Mg ha�1� US$ 232 Mg�1) in a bad yearif CAW is adopted.
Given that the total production costs of CAW areless than that of
CTW (Aryal et al., 2015b; Erenstein and Laxmi,2008), farmers stand
to benefit more under CAW. Therefore, CAWhas both climate-adapted
and economic benefits (in terms of yieldgain and total cost
reduction), implying a win-win situation.Despite these benefits,
farmer uptake of CAW is still relatively slow.Lack of knowledge and
availability of machines required for CAWare two major reasons for
low adoption. In addition, as CAW hasbeen adopted by the farmers in
the study area for last two to fiveyears, it takes time to make
full transition. Targeted policies toenhance farmer access to the
Turbo Happy Seeder are already inplace in Haryana but uptake is
still limited due to the constraints offarmer knowledge and
confidence. Increasing farmers’ knowledge
-
334 J.P. Aryal et al. / Agriculture, Ecosystems and Environment
233 (2016) 325–335
on CAW and building their confidence requires targeted
agricul-tural trainings focused on field demonstrations of
CAW-relatedmachinery, together with regular interactions between
research-ers and farmer societies. Local farmer clubs can play a
crucial rolein initiating such programs.
8.2. Implications for government
Our analysis of long-term weather data (see Section 3) showsthat
untimely excess rainfall events in the wheat season occursonce in 4
years implying that farmers in Haryana may face hugewheat crop
losses due to untimely excess rainfall on a regular basis.As it is
the liability of the government to help farmers when theircrops are
damaged by the extreme climate events, this results in ahuge
financial burden to the government. For example, Haryanagovernment
spent about US$ 175 million on compensation towheat farmers who
suffered from yield losses due to untimely andexcess rainfall in
2014–15 wheat season (The Economic Times,2015). Given this, it
would be worthwhile to look for alternativecrop production systems
that can reduce crop loss under suchextreme climatic events. It may
also be argued that such fundswould have been better utilized in
incentives schemes for theadoption of climate-adapted practices
such as CAW.
In Haryana, 2.54 million ha (Mha) of land was under
wheatproduction in 2014–15 (http://www.agriharyana.nic.in
accessedon 20 November 2015). Haryana contributes almost 11% of
thenational wheat production (HFC, 2012) and so any large scale
lossof wheat production in the state exerts a huge financial burden
onthe government of India. The Haryana state government has set
atarget to increase the area of ZT wheat to one million ha by
2015(HFC, 2012). However, the target is not yet realized and at
present,the area under ZT wheat is approximately 0.3 million ha
(personalcommunication with Dr. Suresh Kumar Gehlawat,
AdditionalDirector Agriculture (General), Department of
Agriculture, Gov-ernment of Haryana). If one million ha of land
under wheat inHaryana had been converted to CAW as targeted, the
state couldhave produced almost 0.66 million Mg (0.66 Mg ha�1�1
Mha)additional wheat in the year 2014–15 (based on the results
inTable 5). This additional production is equivalent to US$
153million when estimated using the minimum support price forwheat
in 2015 in India.
Institutional and policy reforms are required for the
promotionof CAW as a means of adapting to climate change.
Enhancingintegration between state-level plans and local level
climatechange adaptation requirements is crucial for this (Aryal et
al.,2015a). The Haryana state government has already initiated
somepolicy change such as the provision of a subsidy to purchase
themachine (Turbo Happy Seeder) required for CAW. Currently,
thegovernment of Haryana is providing almost a 50% subsidy
tofarmers to purchase Turbo Happy Seeder � INR 50000 (i.e., US$800)
for male farmers and INR 63,000 (i.e., US$ 1008) for thefarmers
with less than 3 acres and female farmers.
Proper evaluation and jointness of various agricultural
subsidyprograms is an important issue in this context. For
example,assuming that the “Turbo Happy Seeder” can serve
approximately50 ha land per wheat season, the Haryana state
requires 20,000units in order to bring one Mha of land under CAW.
If thegovernment bears the full costs of purchasing 20,000 units,
it willcost US$ 40 million in total (one unit cost US$ 2000). Given
that theaverage life of the Happy Seeder machine currently
available on themarket is approximately 9 years, the costs of total
subsidy to thegovernment becomes approximately US$ 4.44 million
yr�1.However, this shift will provide an additional wheat output
of0.66 million Mg (i.e., 0.66 Mg ha�1�1 million ha) annually in a
badyear and 0.41 million Mg (i.e., 0.41 Mg ha�1�1 million ha)
annuallyin a normal year. Therefore, bringing 1 million ha of land
under
CAW will yield an additional economic benefit of US$ 95.12
milliona year (i.e., 0.41 million Mg � US$ 232 Mg�1) in a normal
year andUSD 153.12 million a year (i.e., 0.66 million Mg � US$ 232
Mg�1) ina bad year. This shows that there is a significant economic
gaineven if the government provides a full subsidy to cover the
costs ofpurchasing the Happy Turbo Seeder. However, the mechanism
andpackaging of subsidies/incentives need to be carefully
consideredto reduce any possible negative effects of subsidies and
to optimizeits social benefits. Overall, the economic benefits that
can berealized by investing in the equipment required for CAW is
muchhigher than the total cost of the subsidy. This is a
cruciallyimportant finding in terms of identifying economically
viablestrategies to address climate risks in agriculture.
Strengthening public-private partnership and using localservice
providers as the major information centers can improvefarmer’s
knowledge on CAW. Hence, mobilizing the local farmcooperatives and
local agricultural input/service providers inpartnership with the
local governmental institutions related toagricultural technology
extension services will help to increaseuptake of CA-based
technologies.
9. Conclusions
This study has four main conclusions: i) the magnitude of
yieldloss in wheat during a bad year was less in CAW than
CTWproviding evidence that conservation agriculture-based
practicesin wheat are an effective adaption response to excessive
anduntimely rainfall events that are becoming more frequent
inNorth-West India, ii) As CAW delivers yield advantages in
bothgood and bad years, it is feasible to promote CAW even
withoutsubsidy. However, increasing farmer’s knowledge and
buildingtheir confidence on CAW through regular trainings are
essential,iii) CAW can serve as climate risk adaptation measures
irrespectiveof farm size, iv) analysis of long-term weather data
from Haryanashowed that one in every four year can be bad year in
terms ofextreme rainfall during wheat season, and thus, CAW can be
a cost-effective means of adapting to rainfall variability during
the wheatseason. Given the cost to the government of
compensationpayments to farmers following the aftermath of adverse
climateevents, effective management of subsidy and economic
incentivesto adopt CAW is a critical issue. Lack of timely
availability ofmachines, knowledge of and confidence in CAW are
three majorconstraints to uptake by farmers. Therefore, there is a
need toprovide a series of field-based workshops demonstrating
CA-basedtechnologies to farmers and local service providers.
Trainings andmobilization of local service providers together with
agriculturalresearch institutions and universities can enhance
furtheradoption of CAW whilst raising awareness amongst key
decisionmakers of the evidence that now exists of the associated
economicand adaptative benefits.
Acknowledgements
The authors acknowledge the financial support of CGIARresearch
programs on Climate Change, Agriculture and FoodSecurity (CCAFS)
and CRP Wheat for this study. We also sincerelyacknowledge the
support from farmers of Haryana. Thanks also toall CIMMYT staffs
based at Karnal for their contributions whilecollecting data,
Shakshi Balyan, a research intern in CIMMYT-CCAFS, from Chaudhary
Charan Singh University, Meerut forcollecting data required for the
study and Love Kumar Singh inCIMMYT-CCAFS at Karnal, Haryana for
his support during the fieldsurvey.
-
J.P. Aryal et al. / Agriculture, Ecosystems and Environment 233
(2016) 325–335 335
References
Aryal, J.P., Jat, M.L., Singh, R., Gehlawat, S.K., Agarwal, T.,
2015a. Framework,guidelines and governance for designing local
adaptation plan of action tomainstream climate-smart villages in
India. International Maize and WheatImprovement Center (CIMMYT),
New Delhi, India.
Aryal, J.P., Mehrotra, M.B., Jat, M.L., Sidhu, H.S., 2015b.
Impacts of laser land levelingin rice-wheat systems of the
north-western indo-gangetic plains of India. FoodSecur. 7, 725–738.
doi:http://dx.doi.org/10.1007/s12571-015-0460-y.
Aryal, J.P., Sapkota, T.B., Jat, M.L., Bishnoi, D.K., 2015c.
On-farm economic andenvironmental impact of zero-tillage wheat: a
case of North-West India. Exp.Agric. 51, 1–16.
doi:http://dx.doi.org/10.1017/S001447971400012X.
Cameron, C.A., Trivedi, P.K., 2009. Microeconometrics Using
Stata. Stata Press, Texas,USA.
Erenstein, O., Laxmi, V., 2008. Zero tillage impacts in India’s
rice-wheat systems: areview. Soil Tillage Res. 100, 1–14.
doi:http://dx.doi.org/10.1016/j.still.2008.05.001.
Erenstein, O., Farooq, U., Malik, R.K., Sharif, M., 2008.
On-farm impacts of zero tillagewheat in South Asia’s rice-wheat
systems. Field Crops Res. 105, 240–252.
doi:http://dx.doi.org/10.1016/j.fcr.2007.10.010.
FAO, 2013. Food and agricultural commodities production:
countries by commodity[WWW Document]. URL www.faostat.fao.org
(accessed 5.13.15).
Gathala, M.K., Ladha, J.K., Kumar, V., Saharawat, Y.S., Kumar,
V., Sharma, P.K., Sharma,S., Pathak, H., 2011. Tillage and crop
establishment affects sustainability of SouthAsian rice-wheat
system. Agron. J. 103, 961.
doi:http://dx.doi.org/10.2134/agronj2010.0394.
Global Watch, FAO, 2015. Global information and early warning
system on food andagriculture [WWW Document]. GIEWS Ctry. Br.
India. URL
http://www.fao.org/giews/countrybrief/country/IND/pdf/IND.pdf
(accessed 10.12.15).
Gupta, R., Gopal, R., Jat, M.L., Jat, R.K., Sidhu, H.S., Minhas,
P., Malik, R., 2010. Wheatproductivity in Indo-Gangetic Plains of
India during 2010: terminal heat effectsand mitigation strategies.
Conserv. Agric. 14, 1–3.
HFC, 2012. Working Group Report on Conservation Agriculture for
Sustainable CropProduction in Haryana. Haryana Farmers’ Commission
(HFC). CCS HaryanaAgricultural University, Hisar, India.
Harrington, L., Erenstein, O., 2005. Conservation agriculture
and resourceconserving technologies – a global perspective. In:
Abrol, I.P., Gupta, R.K., Malik,
R.K. (Eds.), Conservation Agriculture – Satus and Prospects.
Centre forAdvancement of Sustainable Agriculture, NASC Complex, New
Delhi, India, pp.1–12.
Jat, R.K., Sapkota, T.B., Singh, R.G., Jat, M.L., Kumar, M.,
Gupta, R.K., 2014. Seven yearsof conservation agriculture in a
rice-wheat rotation of Eastern Gangetic Plains ofSouth Asia: yield
trends and economic profitability. Field Crops Res. 164, 199–210.
doi:http://dx.doi.org/10.1016/j.fcr.2014.04.015.
Khatri-Chhetri, A., Aryal, J.P., Sapkota, T.B., Khurana, R.,
2016. Economic benefits ofclimate-smart agricultural practices to
smallholder farmers in the Indo-Gangetic Plains of India. Curr.
Sci. 110, 1244–1249.
doi:http://dx.doi.org/10.18520/cs/v110/i7/1244-1249.
Kumar, K.K., Kumar, R.K., Ashrit, R.G., Deshpande, N.R., Hansen,
J.W., 2004. Climateimpacts on Indian agriculture. Int. J. Climatol.
24, 1375–1393. doi:http://dx.doi.org/10.1002/joc.1081.
Mas-Colell, A., Whinston, M.D., Green, J.R., 1995. Microeconomic
Theory. OxfordUniversity Press, New York.
Samra, J.S., Singh, G., 2004. Heat Wave of March 2004: Impact on
Agriculture. IndianCouncil of Agricultural Research, New Delhi,
India.
Sapkota, T.B., Jat, M.L., Aryal, J.P., Jat, R.K.,
Khatri-Chhetri, A., 2015. Climate changeadaptation, greenhouse gas
mitigation and economic profitability ofconservation agriculuture:
some examples from cereal systems of Indo-Gangetic Plains. J.
Integr. Agric. 18, 1524–1533.
doi:http://dx.doi.org/10.1016/S2095-3119(15)61093-0.
Shiferaw, B., Smale, M., Braun, H.-J., Duveiller, E., Reynolds,
M., Muricho, G., 2013.Crops that feed the world 10. Past successes
and future challenges to the roleplayed by wheat in global food
security. Food Secur. 5, 291–317.
doi:http://dx.doi.org/10.1007/s12571-013-0263-y.
Sidhu, H.S., Humphreys, E., Dhillon, S.S., Blackwell, J.,
Bector, V., 2007. The HappySeeder enables direct drilling of wheat
into rice stubble. Anim. Prod. Sci. 47,844–854.
Singh, A., Phogat, V.K., Dahiya, R., Batra, S.D., 2014. Impact
of long-term zero tillwheat on soil physical properties and wheat
productivity under rice-wheatcropping system. Soil Tillage Res.
140, 98–105. doi:http://dx.doi.org/10.1016/j.still.2014.03.002.
The Economic Times, 2015. Haryana government announces Rs 1092
crore farmercompensation. Chandigarh, India.
http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0005http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0005http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0005http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0005http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0010http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0010http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0010http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0015http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0015http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0015http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0020http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0020http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0025http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0025http://dx.doi.org/10.1016/j.still.2008.05.001http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0030http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0030http://dx.doi.org/10.1016/j.fcr.2007.10.010http://www.faostat.fao.orghttp://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0040http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0040http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0040http://dx.doi.org/10.2134/agronj2010.0394http://www.fao.org/giews/countrybrief/country/IND/pdf/IND.pdfhttp://www.fao.org/giews/countrybrief/country/IND/pdf/IND.pdfhttp://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0050http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0050http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0050http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0055http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0055http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0055http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0060http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0060http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0060http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0060http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0060http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0065http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0065http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0065http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0065http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0070http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0070http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0070http://dx.doi.org/10.18520/cs/v110/i7/1244-1249http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0075http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0075http://dx.doi.org/10.1002/joc.1081http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0080http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0080http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0085http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0085http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0090http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0090http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0090http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0090http://dx.doi.org/10.1016/S2095-3119(15)61093-0http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0095http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0095http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0095http://dx.doi.org/10.1007/s12571-013-0263-yhttp://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0100http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0100http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0100http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0105http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0105http://refhub.elsevier.com/S0167-8809(16)30464-9/sbref0105http://dx.doi.org/10.1016/j.still.2014.03.002
Conservation agriculture-based wheat production better copes
with extreme climate events than conventional tillage-based s...1
Introduction2 Effects of excess rainfall on wheat production3
Untimely rainfall and wheat crop loss in Haryana4 Study area and
data4.1 Descriptive statistics of the study households
5 Empirical framework of the study5.1 Analysis of major factors
affecting wheat yield under normal and bad years5.2 To test the
yield difference between CA-based and conventional tillage-based
wheat production systems5.3 To test the yield variation across
different farm sizes
6 Results6.1 Factors determining yield difference in normal and
bad year6.2 Yield difference between CAW and CTW6.3 Yield variation
across different farm sizes6.4 Farmers’ perception on why CAW
performs better than CTW under variable and untimely rainfall
conditions
7 Constraints to adoption of CAW8 Discussions and policy
implications8.1 Implications for farmers8.2 Implications for
government
9 ConclusionsAcknowledgementsReferences