Pakistan J. Agric. Res. Vol. 28 No.

1, 2015

CLIMATE VARIABILITY IMPACTS ON RICE CROP PRODUCTION IN

PAKISTAN

Usman Shakoor*, Abdul Saboor*, Irfan Baig*, Anila Afzal**

and Abdul Rahman***

ABSTRACT:- The climate variability has affected the agriculture

production all over the globe. This concern has motivated important
changes in the field of research during the last decade. Climate changes are
believed to have declining effects towards crop production in Pakistan.
This study carries an empirical investigation of the effects of climate
change on rice crop of Pakistan by employing Vector Auto Regression (VAR)
model. Annual seasonal data of the climatic variables from 1980 to 2013
has been used. Results confirmed that rising mean maximum temperature
would lead to reduction in rice production while increase in mean
minimum temperature would be advantageous towards rice production.
Variation in mean minimum temperature brought about seven percent
increase in rice productivity as shown by Variance Decomposition. Mean
precipitation and mean temperature would increase rice production but
simulations scenarios for 2030 confirmed that much increase in rainfall
and mean temperature in long run will negatively affect rice production in
future. It is therefore important to follow adequate policy action to
safeguard crop productions from disastrous effects. Development of
varieties resistant to high temperatures as well as droughts will definitely
enhance resilience of rice crop in Pakistan.

Key Words: Rice Crop; Agriculture; Climate Change; Temperature;

Precipitation; Vector Auto Regression; Pakistan.

INTRODUCTION varying rainfall intensities will be

witnessed around the globe, hence
The climate change issue has translating their effects towards low
been more frightening towards the agricultural output. Climate change is
healthier expansion of both socio- affecting crop yields in many low
economic and agriculture activities of income countries where climate is a
any country (Adejuwon, 2004). key determinant of agriculture
Variation in local climate will force productivity ( Apata, 2010).
people's decision with penalties for Previous evidence has proved that
their economic, personal and social the current global warming has
environment ultimately affecting their turned down agriculture production,
livelihood (UNFCC, 2007). The which lead to a decrease in food
scientific literature showed that in production (Kurukulasuriya and
next decade high temperatures and Mendelsohn, 2008; IISD, 2009).
* Department of Economics and Agric. Economics, PMAS-Arid Agriculture University, Rawalpindi, Pakistan.
** Department of Sociology, PMAS-Arid Agriculture University Rawalpindi, Pakistan.
*** Sustainable Development Policy Institute, Islamabad, Pakistan.
Corresponding author: economistusman@gmail.com

19
 USMAN SHAKOOR ET AL.

Average global temperature for agriculture production. Pakistan's

previous hundred years has agriculture has witnessed number of
o
augmented approximately by 0.8 C mayhems since last 50 years. Some
whereas changes in precipitation years witnessed slow growth, while
levels are not uniform. In humid and the others have recorded healthier
sub humid areas, monsoon rainfalls expansion. This sector is under the
will increase, whereas coastal and threat of climate change as drop in crop
hyper arid areas will encounter yields are due to variations in climate
decrease in winter and summer parameters. At the end of this century,
rainfalls. Sea levels will rise upto 1 to 3 rice crop is expected to encounter
mm per year. These dangerous 15-20% reduction due to climate
changes have emphasized the change which is twice of the impact
importance of mitigation and on wheat crop (Iqbal et al., 2009).
adaptation strategies for the More than 90% of the total
developing economies (IPCC, 2014). acreage of rice is in Punjab province.
Agriculture sector is a direct In Sindh province, primary rice
climate dependent and acts as a vital growing districts are Larkana and
part of the economic activity in Jacobabad and the ad joint district of
developing economies. Climate Nasirabad of Balochistan province. In
change is likely to produce significant Punjab, Gujranwala, Sheikhupura
increase in amount of hydro events and Sialkot districts are primary
and is also expected to bring changes producer of rice. Shikarpur, Dadu,
in temperature to extreme in the Thatta, Badin in Sindh and Okara,
current century (Cline, 2008). Gujrat, Sargodha, Lahore, Kasur and
Increase in temperatures is supposed Sahiwal districts of Punjab are
to produce more floods which will secondary rice growing districts
damage crop yields. Favorable (Abedullah and Mushtaq, 2007).
impacts can be traced out for Major crops of Pakistan are
relatively cooler growing season while affected by changing climate.
unpleasant impacts will be recorded Siddiqui et al. (2012) found the
for regions that have already susceptibility of four major crops of
encountered increasing high tem- Punjab, Pakistan, when they were
peratures (Mendelsohn, 2007). exposed to climatic changes.
Pakistan is situated in the sub- Temperature rise was found
tropical region where climate is important for rice production at first,
temperate. Summer and winter but when temperature increases
seasons are clearly defined and these beyond certain level, the increase
two seasons are imperative for the became destructive towards its
most advantageous crop production. production. Similarly, Mahmood et
Rice is the second staple food of al. (2012) evaluated the effects of
Pakistan after wheat and has an increasing temperature and sugges-
important share in the exports. ted that increase in temperature by
o o
Country's economic performance 1.5 - 3 C will enhance the production
primarily depends on the agriculture while rainfall will hurt rice crop.
production, so a minor deviation in Ajetomobi et al. (2010) concluded
the climate variables may bring that increase in temperature compo-
significant alteration in the overall nent has minimized net revenue of

20
 VARIABILITY IMPACTS ON RICE CROP PRODUCTION

dry land rice farms in Nigeria. Kayam doesn't contain exogenous variable in
et al. (2000) found that drop in the model. Sims (1980) provided VAR
rainfall will reduce wheat yield in model on the foundation of true
Turkey. Peng et al. (2004) observed simultaneity among the exogenous
very close connection between rice and endogenous variables.
yield and average minimum temp- A VAR method offers compre-
erature. Rice production is exhausted hensive results and also decomposes
by 10% when minimum temperature the individual effect which in reply
o
is increased by 1 C. offers the individual impact of the
Climate change is being coun- climate variable on rice production in
tered all over the world and also port- the country. The impulse response
raying its physical impacts every- function is employed to confirm for
where. It is therefore necessary to the shocks in the variables and
evaluate the impact of these changes ultimately to see their impacts on the
non crop sector of Pakistan. Being the explanatory variables. Assessments
cash crop of Pakistan, the impact of made from VAR model are easy
climate change on its production has because Ordinay Least Square (OLS)
been assessed in this study with the method is used (Gujarati and Porter,
objective that empirical evaluation of 2009; Janjua et al., 2010).
the impact of temperature and precipi- VAR model in the matrix form is
tation changes (chief climate parameter) as follows:
on rice production in Pakistan. Zt - 1 2 p Zt-1 et
The study also intends to inspect Zt-1 0 + I 0 0 + Zt-2 + 0
the role of a number of other variables 0
on rice production and will proposes Zt-k+1 0 Zt+k 0
0 I 0
some policy measures to arrest
climate change effects on rice In form of equation the above
production in Pakistan. model can be written as follows:
Zt = + 1Zt-1 + ---- + k Zt-k + et
MATERIALS AND METHOD or
(L)Zt = + et
Vector Auto Regression (VAR) where,
Model (L) = Matrix of polynomial in lag
It is functioned to grab the impact operator.
of all climatic and non-climatic
variables in the model. At first, VAR Specification of the Model
model was operated in macro- General equation including
economics (Janjua et al., 2010). Sims climatic variables and non-climatic
(1980) introduced the VAR model. It variables of the study is:
ostensibly looks like simultaneous Rice production = f (mean temperature,
equation models wherein a number mean min. temperature, mean max.
of endogenous variables are temperature, mean precipitation,
simultaneously considered and each cultivated area, fertilized used, credit
endogenous variable is elucidated by used, water availability).
its past values, and past values of all
other endogenous variables, available Econometric Form of the Model
in the model. Generally VAR model PD = 1 + 2 AC + 3 AVT + 4 CD +

21
 USMAN SHAKOOR ET AL.

5 FR + 6 MINTEM + 7 evaluated by ADF Test.

MXTEM + 7 RAIN + 8 WA + I
where, Lag Selection Criteria for VAR
PD = Production of rice crop Model
AC = Cultivated area Values of Akaike Information
AVT = Mean temperature Criterion (AIC) and Schwarz Criterion
CD = Agricultural loan (SC) of the model enabled to work
FR = Fertilizer used with a VAR model of lag two. The AIC
MINTEM = Mean min. temperature and SWC value for present model are
MXTEM = Mean max. temperature 15.37 and 16.25, respectively, which
RAIN = Mean precipitation is lower than the other lag models
WA = Water availability tested, therefore, this model was
selected. Thus two lag VAR model
Variables and Data Requirements estimated the dynamics of rice
Rice production data was collected production in Pakistan.
from different issues of Agriculture
Statistics of Pakistan and Economic VAR Model
Survey of Pakistan. Data was The results of VAR model
expressed in thousand ton. Data estimation revealed that though t-
regarding the climatic variables i.e statistics of some variables are
mean temperature , mean maximum significant while some are non-
temperature, mean minimum tempe- significant at conventional level of
rature, mean precipitaion was collec- significance however F-statistic value
ted from Pakistan Meterological Dep- of the model is very high and also
artment. Cultivated area under rice statistically significant so the model
crop, fertilizer used, water availability is said to a best fit and this higher
and agriculture loan are the other value has made all the lag terms
explainatory variables used in the significant (Table 1). The coefficient of
2
study. The data of these variables was determination R value is 0.93 and
also extracted from different editions adjusted coefficient of determination
2
of Economic Surveys of Pakistan. R value is 0.83. Both the values are
lying between 0 and 1 and also both
RESULTS AND DISCUSSION the values are very high which also
shows the goodness of the fit of the
overall model.
Vector Auto Regression (VAR)
requires the testing the stationarity of
the variables. For fulfilling this Cholskey Impulse Response
requirement Augmented Dickey Function
Fuller (ADF) test was used (Dickey Impulse response function traces
and Fuller, 1981). Mean temperature, the effect of one time shock on one of
mean min. temperature, mean max. the innovation on current and future
temperature and mean precipitation values of endogenous variable. The
were stationary at level, however impulse response functions of the
fertilizer used, rice production, mean temperature and rice produc-
cultivated area under rice crop, water tion demonstrated that a unit shock
availability and fertilizer used were in the mean temperature will die out
st
stationary at 1 difference as in the period three and remains

22
 VARIABILITY IMPACTS ON RICE CROP PRODUCTION

Table 1. Results of VAR equation

Parameter PD AVT MINTEM MXTEM RAIN WA

PD(-1) 0.216781 0.0000729 0.000491 -0.000564 -0.000403 -0.00284
PD(-2) -0.082455 -0.000510 -0.000477 -0.000442 0.002418 -0.002950
AVT(-1) -201.8721 0.108040 -0.203063 0.510892 -44.29020 2.281388
AVT(-2) -1645.077 -0.378783 -0.345236 0.001458 -31.44283 12.25208
MINTEM(-1) 317.5305 0.692477 -0.101070 1.526184 -21.54851 1.667416
MINTEM(-2) 1330.958 0.182666 0.111873 -0.008025 9.597249 -1.952683
MXTEM(-1) -674.2208 -0.440595 0.081354 -1.119414 34.78596 0.315091
MXTEM(-2) 408.6413 0.043904 -0.234017 -0.013406 8.471804 -3.341756
RAIN(-1) -29.48764 -0.007880 -0.002756 -0.018194 -0.205144 0.021240
RAIN(-2) -11.53409 -0.000556 -0.001531 -0.000759 -0.261167 0.078985
WA(-1) 2.271125 -0.022950 -0.008477 -0.035141 1.423450 0.325274
WA(-2) 38.89460 -0.004012 -0.025723 -0.001506 -0.358144 0.311420

R-squared 0.934591 0.495875 0.856136 0.489795 0.514129 0.987754

Adj. R Squard 0.836477 -0.260312 0.640341 -0.275514 -0.214678 0.969385
F-statistics 9.525573 0.655757 3.967350 0.639997 0.705439 53.77319

Akaike AIC 15.37221 1.334261 0.432374 2.226120 9.034243 4.591039

Schwarz 16.25110 2.213156 1.311269 3.105016 9.913139 5.469935

C 19738.59 27.63125 42.88772 34.50515 507.0972 -213.5823

VAR model estimation results to other variables: Agricultural loan (CD); fertilizers used (FR), cultivated area (AC) are not
presented
negative till period five (Table 2). After rature will also becomes negative in
period three it again becomes positive second period and a single unit shock
and remains positive till the tenth in the mean maximum temperature
period and effects rice production will die out in period two. Non-
positively. From the impulse respon- climatic variables of the study i.e.,
se function of the mean precipitation cultivated land under rice crop,
it is evidently explicable that a single fertilizer used and water availability
standard deviation shock in the mean will become negative in period eight,
rainfall will become negative in period two and three, respectively, while one
two and again becomes positive in unit shock in credit availability will
period three and four. The impulse die out in period seven.
response functions of mean mini-
mum temperature and the rice Variance Decomposition
production made it clear that a unit The intention of variance decom-
shock in the mean minimum tempe- position is to break up the variations

23
 USMAN SHAKOOR ET AL.

Table 2. Impulse Response Function

Period PD AVT MINTEM MXTEM RAIN WA

1 458.8860 0.000000 0.000000 0.000000 0.000000 0.000000

2 193.0700 94.80015 -119.7588 -94.56199 -120.0063 4.095417

3 320.1524 -228.0404 113.0276 42.10135 55.25553 11.01197

4 219.9906 -198.3989 -268.3817 -43.36467 0.888962 -62.72556

5 581.8444 -81.63539 -97.72670 -117.8663 -115.7066 -79.62272

6 339.7266 89.31047 -205.1776 -13.24842 -78.77917 -7.406293

7 353.1206 0.307590 -6.161055 -81.02423 -36.40017 12.07972

8 177.7763 69.82375 -54.14878 90.58391 -25.21080 26.39945

9 102.9457 141.5578 62.01071 -58.71410 -78.59029 65.51808

10 -92.74458 210.5983 87.18666 -8.307131 9.415443 130.6659
Results of other variables, namely credit disbursed (CD), fertilizers used (FR), cultivated area (AC) are not reported

on the endogenous variables in to where it has negative effect on rice

component shocks to the VAR. A production. The highest of the
variation of about seven percent in variation explained in rice production
period ten showed positive effect of due to fertilizer used was about 7% in
mean temperature (Table 3). For mean period five where fertilizer off take
maximum temperature about 2% positively affected rice production. So it
variation produced was in period nine is clear that increasing application of
while in period ten it showed about fertilizer in long run will decrease rice
1.7% variation in rice production, production. Credit availability and
showing the negative effect of area under rice cultivation showed
increasing maximum temperature. about 22% and 15% variations in rice
Variation in mean precipitation in production, respectively.
period ten was about 2% in rice
production. The highest variation due Simulations Scenarios for Year
to mean precipitation was appro- 2030
ximately 3% in period three showing Different simulation scenarios
positive effect of mean precipitation on measured for 2030 have been
rice production. Mean minimum analyzed. In first scenario it was
temperature showed a variation of assumed that temperature increased
o o
approximately 6% in period ten from 2 C to 4 C, in second scenario
o
showing a positive effect on rice temperature increased from 4 C to
o
production as confirmed from impulse 5 C from the base year was taken. The
response function. Variation in water effect was measured in percentage
availability of approximately 1.3% in change in rice production. In scenario
rice production, explaining the positive three, an increase in mean precipi-
effect of time availability of water for tation from 5% to 10% was considered,
rice crop. Variation in fertilizer used while in fourth scenario, an increase
was approximately 5% in period ten from 10% to 15% in rainfall from the

24
 VARIABILITY IMPACTS ON RICE CROP PRODUCTION

Table 3. Variance Decomposition

Period S.E. PD AVT MINTEM MXTEM RAIN WA

1 458.8860 100.0000 0.000000 0.000000 0.000000 0.000000 0.000000

2 583.2014 72.87123 2.642293 4.216746 2.629033 4.234196 0.004931
3 810.7162 53.30457 9.279343 4.125819 1.630174 2.655671 0.021002
4 922.3794 46.86807 11.79519 11.65350 1.480398 2.051694 0.478680
5 1250.042 47.18329 6.848556 6.956112 1.695083 1.973850 0.666343
6 1400.220 43.49157 5.865108 7.691174 1.359928 1.889694 0.533871
7 1457.851 45.98805 5.410570 7.096899 1.563425 1.805586 0.499362
8 1477.386 46.22787 5.491795 7.044791 1.898288 1.787271 0.518173
9 1522.463 43.98820 6.035927 6.799702 1.936271 1.949471 0.673139
10 1606.090 39.86009 7.143098 6.404720 1.742558 1.755180 1.266753
Results of other variables i.e. credit disbursed (CD), fertilizers used (FR), cultivated area (AC) are not reported

base year was considered. pitation by 10% to 15% from base

In first scenario temperature in- year will decrease rice production by
o o
crease from 2 to 4 degree till 2030 approximately 0.6%; this decrease is
will decrease rice production by 8% not significantly high but further
approximately. In second scenario increase will decrease rice production
o
when temperature increase from 4 to more severely. Mean minimum tem-
o
5 , percentage decrease in rice perature will effect positively the rice
production will increase and will be production; however increase in
about 11.5%. In rainfall scenarios mean maximum temperature will
when rainfall increased from 5% to effect rice production negatively.
10%, rice production will decrease by Jayanta (2011), portrayed negative
only 0.3 % and when it increased from effects of increasing maximum tem-
10% to 15% decrease was about perature on rice yields in Bangladesh.
0.59%. It is clear that too much Impulse response and variance
increase in temperature and also pre- decomposition also showed that
cipitation will affect the rice production variation due to mean temperature
negatively. towards rice production was about
VAR Model results and climate 7% in period ten, which highlighted
change scenarios showed that the positive effects but in long run
increase in mean precipitation will scenario increa dangerous for rice
last positive effects towards rice production. For 2030 it was revealed
production in Pakistan. Subash and that if mean temperature increases
o o
Mohan (2010) also proposed that from 4 C to 5 C, the rice production
rainfall will increase rice production. would decrease approximately 11.5%
In long run when mean precipitation from the base year i.e., 20se in mean
will increase too much, the negative temperature will be highly13. Subra-
effects would be encountered on rice hmanyam et al. (2009) and Ajetomobi et
production. Simulations for 2030 al. (2010) presented the negative effects
suggested that increase in preci- of increasing mean temperature on rice

25
 USMAN SHAKOOR ET AL.

production. The non-climatic variables pacts and adaptation to climate

will also affect the rice production as change. Obafemi Awolowo Uni-
analysis provided us that too much versity, I1e-Ife, Nigeria. p. 2-8.
increase in fertilizer application would Ajetomobi, J.O., A. Abidun and R.M.
also disturb rice production, while Hassan. 2010. Economic impact
timely availability of the water will of climate change on irrigated rice
significantly add towards rice crop agriculture in Nigeria. Contri-
production. buted paper presented at the
It is thus concluded that this Joint 3rd African Association of
evolving risk of climatic change will Agricultural Economists (AAAE)
stress the production of rice in and 48th Agricultural Econo-
Pakistan. Rainfall and mean tempe- mists Association of South Africa
rature increase would be beneficial (AEASA) Conference, Cape Town,
for healthier rice production but will S. Africa, Sept. 19-23.
produce negative effects if these Apata, T.G. 2010. Effects of global
climatic variables are increased too climate change on Nigerian
much in future decades as evident agriculture: An empirical
from the simulation scenarios. Inc- analysis. CBN J. Appl. Statistics,
rease in mean maximum temperature 2(1): 45-60.
will impair rice production while Cline, W.R. 2008. Global Warming
increase in minimum temperature and Agriculture. Finance and
will raise rice production. Varieties Development. p. 23-27.
which are tolerant to high tempe- David, A.D. and W.A. Fuller. 1981.
rature and drought should be Likelihood ratio Statistics with
developed so that losses could be autoregression time series with a
avoided. The temperature component unit root. Econometrica, 49(4):
may shorten the growth periods; 1057-1072.
therefore the cultivating time should Gujarati, D.N. and D. Porter. 2009.
be adjusted accordingly. A Climate Basic Econometrics. Mc Graw-
Smart rice production system should Hill, N.Y. 5th edn. p. 711-737.
be maintained which is backed by IISD (International Institute for
comprehensive extension system. Sustainable Development). 2009.
Community based adaptation to
LITERATURE CITED climate change Bulletin. A
Summary of the Second Interna-
Abedullah, K.S. and K. Mushtaq. tional Workshop on Community
2007. Analysis of technical Based Adaptation to Climate
efficiency of rice production in Change. 135(2): 1-23.
Punjab (Pakistan): Implications IPCC (Intergovernmental Panel on
for future investment strategies. Climate Change) 2014. Summary
Pakistan Econ. and Social Rev. for the Policymakers. In:
45(2): 231-244. Edenhofer, O., R. Pichs-Madruga,
Adejuwon, S.A. 2004. Impacts of Y. Sokona, E. Farahani, S. Kadner,
climate variability and climate K. Seyboth, A. Adler, I. Baum. S.
change on crop yield in Nigeria. Brunner, P. Eickemeier, B.
Paper presented at Stakeholders' Krieman, J. Savolainen, S.
workshop on assessment of im- Schlomer, C. Von Stechow, T.

26
 VARIABILITY IMPACTS ON RICE CROP PRODUCTION

Zwickel and J.C. Minx (eds.). cropping system of Punjab

Climate Change. Mitigation of Province. J. Anim. Pl. Sci. 22(4):
Climate Change. Contribution of 993-997.
working Group III to the Fifth Mendelsohn, R. 2007. Past climate
Assessment Report of IPCC. change impacts on agriculture.
Cambridge Uni-versity Press, Chapter 60, Handbook of Agri-
Cambridge, United Kingdom and cultural Economics. In: Evenson
New York, NY, USA. 31 p. R. and P. Pingali (eds.) North
Iqbal, M.M., M.A. Goheer and A.M. Holland: Elsevier. 3: 3009-3031.
Khan. 2009. Climate-change Peng, S., J. Huang, J.E. Sheehy, R.C.
aspersions on food security of Laza, R.M. Visperas, X. Zhong,
Pakistan. Science Vision, 15(1): G.S. Centeno, G.S. Khush and
15-23. K.G. Cassman. 2004. Rice yields
Janjua, P.Z., G. Samad and N.U. Khan. decline with higher night
2010. Impact of Climate Change temperature from global
on wheat production: A Case warming. Proc. Nat. Acad. Sci.
Study of Pakistan. The Pakistan 101(27): 9971-9975.
Develop. Rev. 49(4): 799-822. Siddiqui, R., G. Samad, M. Nasir and
Jayanta, K.B. 2011. Climate change H.H. Jalil. 2012. The Impact of
impacts on rice production in Climate Change on Major
Bangladesh: Results from a Agricultural Crops: Evidence from
model. Centre for Research and Punjab, Pakistan. The Pakistan
Action on Development. p. 6-24. Develop. Rev. 51(4): 261-276.
Kayam, Y., U. Ozsoy, J. Lomas, O. Sims, C.A. 1980. Macroeconomics
Oden, M. Mandel and M. Gurbuz. and reality. Econometrica, 48(1):
2000. The impact of climatic 1-48.
change on wheat production of Subash, N. and H.R. Mohan. 2010.
Aegean region in Turkey: The Assessment of the influence of
effect of a reduction of rainfall monsoon rain on rainy season rice
and the increase of temperature (Oryza sativa) productivity over
on wheat yields.(Available at major rice growing states. Indian
http://www,fao.org/sd/climagri J. Agric Sci. 80(7): 606-615.
med/pdf/ws01_26.pdf). Subrahmanyan, D.R., D.R. Rao,
Kurukulasuriya, P. and R. Men- P.M.V. Reddy and S.R. Voleti.
delsohn. 2008. A Ricardian 2009. Climate change and its
analysis of the impact of climate impact on rice. Rice Knowledge
change on African crop land. Afri. Management Portal (RKMP)
J. Agric. Resour. Econ. 2(1): 1-23. (Available at http://www.
Mahmood, N., B. Ahmad, S. Hassan rkmp.co.in)
and K. Bakhsh. 2012. Impact of UNFCC. 2007. Climate change
temperature and preciptation on impact, vulnerability and adap-
rice productivity in rice-wheat tation in developing countries. J.

(Received December 2014 and Accepted February 2015)

27