Content uploaded by Nazim Ali
Author content
All content in this area was uploaded by Nazim Ali on Dec 29, 2014
Content may be subject to copyright.
Climatic Change (2009) 94:123–142
DOI 10.1007/s10584-009-9559-5
Vulnerability and adaptability of wheat production
in different climatic zones of Pakistan under climate
change scenarios
Humaira Sultana ·Nazim Ali ·M. Mohsin Iqbal ·
Arshad M. Khan
Received: 22 February 2008 / Accepted: 29 December 2008 / Published online: 24 February 2009
© Springer Science + Business Media B.V. 2009
Abstract Ten wheat production sites of Pakistan were categorized into four climatic
zones i.e. arid, semi-arid, sub-humid and humid to explore the vulnerability of wheat
production in these zones to climate change using CSM-Cropsim-CERES-Wheat
model. The analysis was based on multi-year (1971–2000) crop model simulation runs
using daily weather series under scenarios of increased temperature and atmospheric
carbon dioxide concentration (CO2) along with two scenarios of water management.
Apart from this, sowing date as an adaptation option to offset the likely impacts
of climate change was also considered. Increase in temperature resulted in yield
declines in arid, semi-arid and sub-humid zone. But the humid zone followed a
positive trend of gain in yield with rise in temperature up to 4◦C. Within a water
regime, increase in CO2concentration from 375 to 550 and 700 ppm will exert
positive effect on gain in wheat yield but this positive effect is significantly variable in
different climatic zones under rainfed conditions than the full irrigation. The highest
response was shown by arid zone followed by semi-arid, sub-humid and humid zones.
But if the current baseline water regimes (i.e. full irrigation in arid and semi-arid
zones and rainfed in sub-humid and humid zones) persist in future, the sub-humid
zone will be most benefited in terms of significantly higher percent gain in yield
by increasing CO2level, mainly because of its rainfed water regime. Within a CO2
level the changes in water supply from rainfed to full irrigation shows an intense
degree of responsiveness in terms of yield gain at 375 ppm CO2level compared to
550 and 700 ppm. Arid and semi-arid zones were more responsive compared to sub-
humid and humid zones. Rise in temperature reduced the length of crop life cycle
in all areas, though at an accelerated rate in the humid zone. These results revealed
that the climatic zones have shown a variable intensity of vulnerability to different
H. Sultana (B)·N. Ali ·M. M. Iqbal ·A. M. Khan
Global Change Impact Studies Centre (GCISC), National Centre for Physics (NCP) Complex,
Quaid-i-Azam University Campus, Islamabad, -44000, Pakistan
e-mail: sultana.humaira@gmail.com
124 Climatic Change (2009) 94:123–142
scenarios of climate change and water management due to their inherent specific and
spatial climatic features. In order to cope with the negative effects of climate change,
alteration in sowing date towards cooler months will be an appropriate response by
the farmers.
1 Introduction
Wheat is the staple food crop of Pakistan which is grown almost all over the country
despite the fact that Pakistan has a highly diversified climate. The inland areas
get as hot as 52◦C while the temperature in the mountains drops to well below
freezing point. Most of the areas in central and southern Pakistan, considered as
granary of the country, are arid whereas the Northern part of the country is humid
except the extreme northern mountains which are dry (Quadir et al. 2004). These
areas have different geographic features that attribute them their specific climatic
characteristics.
The wheat productivity in the country has been increasing as a result of green
revolution during the past four decades. There are, however, increasing concerns that
abiotic and biotic stresses may put strains on sustainable productivity. Among the
abiotic stresses threatening food security in Pakistan as well in South Asia, climate
change is a major factor. That is why; food security will be at top of the agenda in
Asian countries in the near future, mainly due to two reasons; growing population
and many direct and indirect effects of climate change (IPCC 2001).
If average temperature were to increase by 2.5◦C, heat stress would reduce wheat
yields by as much as 60% (Pakistan Country Report 1994). The dependence of
agriculture on climatic factors (light, heat, water) implies that the crop production
will be significantly affected by changes in climate (Parry 1978; Thompson 1975;
World Meteorological Organization 1979). In some regions climate change may lead
to prolonged dry spells or more intensified heat waves, seriously damaging crop
production by causing moisture or thermal stress (Rounsevell et al. 1999 in the IPCC
2001). Climate change would not only increase the water demand for agricultural
crops due to increased evapo-transpiration but would also adversely affect the water
availability for crops (Ullah et al. 2001). It is also well known that water limitation
tends to enhance the positive crop response to elevated CO2,comparedtowell
water conditions (Chaudhuri et al. 1990; Kimball et al. 1995). Because of recent
climatic variations, the arid zone of the country is facing severe scarcity of water,
which has tremendously affected human life, ecology and economic activities. Thus
understanding the possible impacts of climate change are of utmost importance from
the national and regional point of view (Quadir et al. 2004).
The extent and nature of negative impacts induced by climate change can be
managed by effective adaptation. In the context of climate change, adaptation refers
to adjustments in human and natural systems to respond to actual or expected climate
impacts. In general, the preferred adaptation strategies are actions with multiple
economic and environmental benefits for current and future conditions and needs,
based on sound scientific assessment. The range of practices that can be used to adapt
to climate change is diverse, and includes changes in behavior, structural changes,
Climatic Change (2009) 94:123–142 125
policy based responses, technological responses or managerial responses (FAO,
Climate Change Adaptation <http://www.fao.org/climatechange/49371/en/>).
Crop models are tools that can be used to evaluate agricultural production risk as
a function of climatic variability, assess regional yield potential across a wide range of
environmental conditions, and determine suitable management factors for increasing
production (Egli and Bruening 1992; Meinke et al. 1993; Aggarwal and Kalra 1994;
Meinke and Hammer 1995). Models can thus be used to extend research results
both spatially and temporally to estimate the performance of crops under changing
climatic and environmental conditions (Pannkuk et al. 1998). Apart from this they
also have ability to test certain adaptation options in order to offset the impacts of
climate change.
This study was designed with the objective of evaluating the relative sensitivity
of wheat production in different climatic zones of Pakistan under climate change
scenarios of temperature and CO2increase, using simulation modeling. The extent
and nature of negative impacts imposed by climate change can be managed by
effective adaptation. The results of the vulnerability assessment were further used to
evaluate shifts in sowing dates as adaptation option that could reduce the potential
negative effects of climate change assuming that it will be a no-cost decision by the
farmers.
This study will not only help to analyze the comparative advantage of different
climatic zones in the production of wheat crop and allocation of resources at current
and under climate change situations but also to device some coping mechanism to
reduce yield losses.
2 Physiography of Pakistan
Pakistan is geographically situated between 24–37◦N latitudes and 62–75◦E longi-
tude and lies on the western margin of one of the climatic regions of the world, the
monsoon region. It is a land of varied landscapes ranging from perpetually snow
capped peaks of Himalayan Range like the Karakoram, K-2 elevation 28,265 ft.
(8,615 m) to lush green canal irrigated plains and the hot dry deserts of Sindh
and Baluchistan where summer temperatures can exceed 50◦C (Framji et al. 1982).
Pakistan is divided into five physiographic regions, each with a number of sub-
regions: (1) the Himalayas, (2) the Hindukush and the Western Mountains, (3) the
Pothwar Plateau and the Salt Range, (4) the Indus Plain, and (5) the Balochistan
Plateau. In addition, the coastal zone and the offshore Exclusive Economic Zone in
the Arabian Sea can be considered physiographic regions. The climate of Pakistan
varies widely, from temperate in the north to hot and dry tropical in the south. The
country is characterized by wide variations in both temperatures and precipitation.
Temperatures reach as low as –26◦C over the northern mountains, and as high as
52◦C over the central arid plains. The mountainous and sub-mountainous areas of
the northeast receive over 1,700 mm of precipitation annually, in contrast to the arid
plains of southwest Balochistan, which receive an average of only 30 mm per annum.
Pakistan receives rainfall in both the summer and the winter. In general, the Punjab
126 Climatic Change (2009) 94:123–142
and Sindh provinces receive more rainfall in the summer, whereas Balochistan and
NWFP receive more rain in the winter (O’Brien 2000).
3 Methodology
3.1 Wheat production sites/climatic zones
Ten sites (Fig. 1) were selected from the entire Pakistan based on availability of data
and their ability to represent different climatic zones. These sites were categorized
into four climatic zones mainly based on Thornthwaite’s Precipitation Effectiveness
index (Eq. 1) (Thornthwaite 1931) by using climatic data of 30 year (1971–2000).
PE Index =
2000
K=1971
1.65 PkTk+12.20.9(1)
Where, “PE” stands for precipitation effectiveness, “PK” stands for total amount of
precipitation (mm) in kth year and “Tk” stands for average daily temperature of kth
year. Figure 1shows the latitude, longitude, elevation and PE index value of each
selected site.
Apart from PE Index, the horizontal and vertical proximity of sites was also kept
in consideration, because it is possible that two sites having no horizontal and vertical
proximity get grouped into one zone, based on the PE Index value.
The climatic zones identified were arid, semi-arid, sub-humid and humid with their
PE index values of less than 16, 16–31, 32–63, 64–127 respectively. Arid zone is the
driest zone, with 30-year annual average rainfall of 186 mm, while the semi-arid, sub-
humid and humid zones have an annual average rainfall of 516, 1,062 and 1,776 mm
Fig. 1 Pakistan map showing
the location of ten sample
meteorological stations with
their altitude and precipitation
effectiveness index value
respectively
Climatic Change (2009) 94:123–142 127
Arid
-10
0
10
20
30
40
50
Sub-humid
Month
Jan
Feb
Mar
April
May
June
July
Aug
Sep
Oct
Nov
Dec
Tempeature (oC)
-10
0
10
20
30
40
50
Semi-arid
Rainfall (mm)
0
100
200
300
Humid
Jan
Feb
Mar
April
May
June
July
Aug
Sep
Oct
Nov
Dec
0
100
200
300
Tmax Tmin Total rainfall
Fig. 2 The 30-year mean monthly maximum (Tmax) and minimum (Tmin) temperature and rainfall
(1971 to 2000)
respectively. The mean monthly variations in maximum and minimum temperature
and rainfall are shown in Fig. 2. Humid zone has the lowest minimum and maximum
temperatures and the highest rainfall compared to other zones.
3.2 CSM-Cropsim-CERES-Wheat model
In this study, the CSM-Cropsim-CERES-Wheat model (Ritchie et al. 1998; Jones
et al. 2003) that is part of the Decision Support System for Agrotechnology Transfer
(DSSAT) Version 4.0 (Tsuji et al. 1994; Hoogenboom et al. 2004) was used. CERES-
Wheat is a dynamic and mechanistic crop growth model that simulates the duration
of vegetative and reproductive stages, accumulation and partitioning of biomass, and
grain yield for a specific cultivar (Hoogenboom et al. 1994; Ritchie et al. 1998). The
model is capable of simulating the impact of the main environmental factors, such
as weather and soil type, and also can provide farmers with information for their
management decisions (Tsuji et al. 1998).
128 Climatic Change (2009) 94:123–142
3.3 Input data
For genetic characteristics, optimized coefficients of spring wheat (Triticum aestivum
L.) cultivar ‘Inqlab 91’ were used (Iqbal et al. 2008). This cultivar covers about 70%
of the total area under wheat production in Pakistan and has excellent characteristics
for wider adaptation. It is a general purpose variety that is suitable for both early and
late sowing.
For the ten sites (Fig. 2), time series daily data of rainfall and minimum and
maximum temperatures, covering the period 1971–2000, were obtained from Pak-
istan Meteorological Department. Data on solar radiation was computed by using
‘Donatelli and Bellocchi’ (DB) model (Donatelli and Bellocchi 2001)ofRadEst
software v 3.0 (Donatelli et al. 2003) which estimates the daily total radiation using
maximum and minimum temperature. For parameterization of the model, observed
data on solar radiation for some scattered number of years at Islamabad, Faisalabad,
Bahawalpur and Shangla were used. These parameterizations were used in such a
way that 30 year (1971–2000) solar radiation data for Islamabad and Jhelum were
estimated with the help of coefficients optimized from Islamabad site; for Faisalabad
and Sheikhupura the optimized coefficients at Faisalabad were used; for Bahawalpur,
Hyderabad, Jacobabad, and D.I. Khan, the optimized coefficients obtained from
Bahawalpur site were used, and for Murree, the coefficients derived from Shangla
were used. In this way for the parameterization of model a separate representative
site for each climatic zone of available observed data of solar radiation was used.
For all simulation runs, nitrogen dosage was kept constant at 100 Kg ha−1at the
time of sowing. Each site had its own soil information collected from Soil Survey of
Pakistan based on the criteria that the soil series possibly should most extensively
represent the area and it must be suitable for the crop production (land capability
classification not lower than 3).
Sowing dates varied, depending on the climatic conditions of the area, and the time
of harvest of the previous crop. For this purpose, each zone was assigned a different
sowing date based on the dates used by farmers in these areas. Sowing dates were
set as 1st November, 15 November, 30 November and 15 December for humid, sub-
humid, semi-arid and arid zones respectively.
3.4 Model validation
For validation purpose the observed data on two crop components (grain yield
and harvest date) of spring wheat (Triticum aestivum L.) cultivar ‘Inqlab 91’ were
acquired form National Uniform Wheat yield Trials (NUWYT) (Mustafa et al. 1998,
2000,2001). The data were of 3 years (1997–1998, 1999–2000 and 2000–2001) at four
locations Faisalabad, Bahawalpur, D.I. Khan and Islamabad. Grain yield and length
of crop life cycle were simulated by the crop growth model by using of measured
inputs on respective soil, crop management and meteorological data.
Apart from the percentage difference of simulated values from observed values,
the d-index (index of agreement) and RMSE (root mean square error) were used as
statistical criteria for model validation.
Climatic Change (2009) 94:123–142 129
The d-index (Willmott 1982) was computed as
d=1−n
i=1
(Pi −Oi)2/
n
i=1Pi+Oi2
Where Oiand Piare observed and predicted values respectively for the i-th data
pair, Pi=Pi −O)andOi=Oi −O,whereOis the average of the observed. The
value for the d-index ranges from zero to one, for good model performance it should
approach unity.
RMSE (Willmott 1982) was computed as
RMSE =N−1
n
i=1
(Pi −Oi)20.5
Where Nis the number of observed values and Oi and Pi are observed and predicted
values for the i-th data pair.
Simulated grain yield was compared with observed grain yield in Fig. 3a. On the
average, the model yields were overestimated by 6% however, yield was underesti-
mated by 4% and 19% at two points (year 1997–1998 and 2000–2001 at Faisalabad
location); the reason for this underestimation remains unexplained. The value of
RMSE (386 Kg/ha−1) and of d-index (0.90) suggest a reasonable estimate of grain
yield by the model.
Days to maturity (length of crop life cycle) were estimated with the help of
respective sowing and harvesting dates in the respective years and locations. The
length of crop life cycle was underestimated by the model in all the cases (Fig. 3b).
Underestimation was, on average, of 22 days (ranges from 12–31 days) with RMSE of
Observed yield (ton ha-1)
3.0 3.5 4.0 4.5 5.0 5.5 6.0
Simulated yield (ton ha-1)
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
Faisalabad
Bhawalpur
D.I Khan
Islamabad
Observed days to maturity
140 150 160 170 180 190 200
Simulated days to maturity
130
140
150
160
170
180
Faisalabad
Bhawalpur
D.I Khan
Islamabad
(a) (b)
Fig. 3 Scatter plot showing observed and simulated grain yield (a) and length of crop life cycle (b)
at Faisalabad, Bhawalpur, D.I. Khan and Islamabad
130 Climatic Change (2009) 94:123–142
23 days and value of d-index 0.57. This systematic underestimation can be attributed
to the gap between harvest date and physiological maturity date as simulated by
the model. If the observed values are corrected for the systematic error (i.e., the
observed values are lowered by 12 days in order to bring them down from harvest
date to physiological maturity), the RMSE was reduced to 11 days with d-index of 1.
If this is corrected by the factor of 22 (average underestimation between observed
and simulated) then the RMSE was reduced to 5 days with d-index of 1. So, by
introducing the correction factor, the fit between the simulated and observed days
to maturity is generally considered satisfactory.
3.5 Simulation scenarios
The long-term (1971–2000) solar radiation, maximum and minimum temperature
and rainfall at the representative sites were used as baseline climatic data. The
weather series for simulations in the changed climate was obtained by a direct
modification of observed series assuming a uniform change in temperature and CO2
over a study area. For CO2, three scenarios were used: 375 (recent), 550 (double of
pre-industrial CO2concentration) and 700 ppmv (almost 2.5 times the pre-industrial
CO2concentration). For temperature, six scenarios (0–5◦C with steps of +1◦Cper
scenario) were used. Though these are synthetic scenarios but range covers well the
IPCC projections for temperature and CO2change to the end of twenty-first century.
These scenarios also seek to evaluate model’s responses to incremental changes in
temperature and CO2both individually and in combination.
Changes in the hydrological regimes in which crops grow are likely to change with
global warming. Although Global Circulation Models (GCMs) predict an overall
increase in mean global precipitation with an unequal spatial distribution yet in some
areas it will decrease. The crop water regime may further be affected by changes
in seasonal precipitation, within-season pattern of precipitation, and inter-annual
variation of precipitation (O’Brien 2000). Because of this uncertain pattern, two
scenarios of water supply i.e. rainfed (with no artificial supplies of water) and full
irrigation (water supply as per requirement of crop) were considered.
In this way, for climate change impacts, we took four factors (temperature, CO2,
water and climatic zones) to design experiment for simulation runs. Temperature
has six levels (0◦C, 1◦C, 2◦C, 3◦C, 4◦C, and 5◦C), CO2has three levels (375, 550 and
700 ppmv), water has two levels (rainfed and full irrigation) and ten sites, forming a
total of 360 treatments, each treatment was run for 30 years (1971 to 2000) of time
series climatic data.
In order to provide reference points with which to compare both scenarios of
future climate and their modeled effects on crop productivity, climatological baseline
was set at 375 ppm CO2by using climatic baseline data as mentioned above, this
baseline remains same for all the climatic zones. But the second baseline that
represents the present-day levels of management and technology varies in terms of
water regimes and sowing dates in all climatic zones. Arid and semi-arid zones are
dominantly irrigated so full irrigation is assumed as baseline water regime in these
zones, but sub-humid and humid zones are rainfed so, rainfed condition is assumed
as their baseline water regime. Regarding sowing dates; the respective sowing dates
in all zones as mentioned in Section 3.3 are taken as baseline sowing dates.
Climatic Change (2009) 94:123–142 131
Apart from this, alterations in the current baseline sowing dates of the respective
zones was also considered to see the sowing date as an adaptation option to offset
the likely impacts of climate change. For this purpose, each zone was run for six
sowing dates (15-Oct, 1-Nov, 15-Nov, 30-Nov, 15-Dec and 30-Dec), assuming that
in future the baseline water regimes of the respective zones will remain same. For
climate scenarios, apart from baseline climate, two scenarios of climate change were
considered that are 550 ppm CO2concentration with 3◦C rise in temperature and
700 ppm CO2concentration with 5◦C rise in temperature with the assumption that
these might be the most plausible combinations of CO2and temperature increase in
future.
In the Section 4, wherever found necessary, the significance of the results is
supported at medium confidence level (alpha value 0.05) using Analysis of Variance
(ANOVA) and Least Significance Difference (LSD).
4 Results and discussion
4.1 Grain yield of wheat at baseline scenario
In the respective baseline scenarios of the specified zones, average yield was the high-
est in semi-arid zone followed by arid, sub-humid and humid zones respectively. This
trend is also confirmed from the observed statistical yields (Fig. 4). For this purpose
the 12 year (1983–1994) statistically reported yields (GoP 1983–1984-to-1994–1995)
in the same areas (except for Murree, which is replaced with Dir based on their
similarities in climatic features) as used in this study were used keeping in view
simulated obs. stat. obs. NUWYT
Yield (ton ha-1)
0
1
2
3
4
5
6
semi-arid arid sub-humid humid
Fig. 4 In the baseline scenario, grain yield comparisons in simulated, obs. Stat (yields from
agricultural statistics) and obs. NUWYT (yields from National Uniform Wheat Yield Trial) sources
in semi-arid, arid, sub-humid and humid zones
132 Climatic Change (2009) 94:123–142
their water supply mode. Though the observed yields didn’t match simulated yields
in magnitude but the trend is similar for all zones in both observed and simulated
yields.
Apart from this, the NUWYT data (as described in Section 3.4) were also grouped
into zones. We don’t have yield data of all sites under study in the experiment, but
only of four locations in which two locations (Bahawalpur and D.I. Khan) represent
the arid zone, Faisalabad represent the semi-arid zone and Islamabad represent the
sub-humid zone while no data for humid zone was available. These experimental
yields also have the same pattern of change in yield with respect to zones as those of
simulated (Fig. 4).
4.2 Effect of temperature increase on grain yield of wheat
Compared to the respective baseline scenario of each zone, yield declined in all
the zones with each unit rise in temperature except in the humid zone (Fig. 5).
The magnitude of decline was the highest in sub-humid zone followed by semi-
arid and arid zones (Fig. 6).The percentage decline in yield in arid and semi-arid
zones is found to be statistically insignificant. But a significant difference exists in
sub-humid zone from arid zone. The higher percentage decline in sub-humid zone
compared to arid and semi-arid zone can be attributed to their different water
regimes, as the sub-humid zone is rainfed while arid and semi-arid zones are run
under irrigated conditions. To cross test this reasoning, these three zones when run
under homogeneous water regimes i.e. either rainfed or full irrigation, the percentage
change in yield with increasing temperature became insignificant among these zones.
Temperature change (°C)
031425
Yield (ton ha-1)
2
3
4
5
Arid Semi-arid Sub-humid Humid
Fig. 5 Trends in simulated wheat yields with temperature changes at 375 ppm CO2concentration in
arid, semi-arid, sub-humid and humid zones
Climatic Change (2009) 94:123–142 133
13245
-50
-40
-30
-20
-10
0
10
20
30
Percent change
Temperature change (oC)
Arid Semi-arid Sub-humid Humid
Fig. 6 Percentage change in yield from base yield in response to temperature changes at 375 ppm
CO2concentration in arid, semi-arid, sub-humid and humid zones
The decrease in yield can largely be attributed to shortened growth periods as
discussed in the following Section 4.4. This is because of accelerated phenology under
increased air temperatures. Grain filling in wheat is a critical stage for high temper-
ature injury (Johnson and Kanemasu 1983). If the crop observes high temperature
at reproductive stages, it hampers normal grain development and leads to shriveled
grain and subsequently the drastic yield losses (Sultana 2003). Results are also in
accordance with the findings of Bender et al. (1999), who reported that a constant 1◦C
increase in temperature over the whole season would decrease yields by 6–10% due
to shorter duration of crop growth. Qureshi and Iglesias (1994) have also reported
based on their modeling work that higher temperatures will drastically reduce yields
in Pakistan. This suggests that these zones are vulnerable to increase in temperature,
especially given their existing water shortages and the high temperatures that already
approach tolerance limits (CGIAR 2004–2005; Parry et al. 1988).
In the scenario of temperature increase, humid zone is an optimistic zone. In this
zone beneficial effects are likely to ensue with higher temperatures. Yield follows a
positive trend of gain with rise in temperature up to 4◦C(Fig.6). On the average,
there was 20% gain in yield from base yield by increasing temperature up to 5◦C
(Fig. 4). Though the length of crop life cycle also gets reduced in this zone but this
shortening is beneficial. Husssain and Mudasser (2007) have also reported based
on econometric analysis that shortening of the length of crop life cycle could be
beneficial for the mountain areas above 1,500 m altitude and they may therefore
expect yield increases because warmer temperatures will make it more likely that
the crop will mature earlier.
In this zone low temperature (cold stress) is a constraint for wheat crop. The spring
wheat crop grown in winter season is subject to cold stress. So, rise in temperature is
134 Climatic Change (2009) 94:123–142
in favor of this zone in terms of yield gain. A similar trend was simulated by Sultana
et al. (2005), who studied wheat crop grown in sub-humid and humid zones and
reported inverted yield trends in these zones. In spite of this positive yield trend
with increase in temperature, the humid zone has low average base yields compared
to the average base yields of arid, semi-arid and sub-humid zones. At temperature
increase of 1◦C, it has approximately approached the yield level of sub-humid zone
but lower than the yields of arid and semi-arid zones.
The comparatively lower yield in the humid zone may be attributed also to non
availability of cold tolerant winter or facultative cultivars in this zone (Hashmi and
Shafiullah 2003). The expected future increases in temperature caused by global
warming would, however, render the varieties of arid and semi-arid zones unsuitable
there and these could be introduced in the humid zone. Rosenzweig (1985) found
that for United States the major effect of climate change would be regional shifts in
the use of wheat cultivars.
4.3 Interactive effects of increase in temperature and CO2along with change in
water supply on grain yield of wheat
In the scenario of baseline water regimes, increased concentrations of CO2from
the baseline level of 375 to 550 and 700 ppm exerted substantial positive effect on
wheat yield in all climatic zones. This positive impact has compensated the negative
impact arising due to increased temperatures up to 2–3◦Cby550ppmandupto3–
4◦C by 700 ppm in arid, semi-arid and sub-humid zones (Fig. 7a, b). In these zones,
increase in temperature beyond 2–4◦C could not sustain baseline yields and nullified
the beneficial effects of enhanced carbon dioxide concentration. Kalra et al. (2003)
Temperature change (°C)
012345 012345
Yield (ton ha-1)
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
Arid Semi-arid Sub-humid Humid
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
(a) (b)
Fig. 7 Trends in simulated wheat yields with changes in temperature at 550 (a) and 700 ppm (b)
CO2concentration in arid, semi-arid, sub-humid and humid zones. (Dotted line with same symbol
represent constant base yield of respective zone)
Climatic Change (2009) 94:123–142 135
also reported that increasing temperature nullifies the beneficial effects of enhanced
carbon.
In humid zone, the pattern of yield gain with rise in CO2concentration (550 and
700 ppm) was similar to the pattern of yield gain at 375 ppm CO2. Changing CO2
concentration level from 375 to 550 and 700 ppm has not changed the trend but
only affected the magnitude of yield positively at varying temperature regimes. It
implies that at the CO2concentration level of 375, 550 or 700 ppm, a 4◦C increase
in temperature in humid zone is the optimal one where maximum yield could be
obtained. In the arid, semi-arid and sub-humid zones, the temperature optima shift
slightly upwards with increasing CO2concentration from 550 to 700 ppm.
Within a water regime, increase in CO2concentration from 375 to 550 and
700 ppm exerted positive effect on gain in wheat yield but this positive effect is
significantly variable in different climatic zones under rainfed conditions. The highest
response in terms of percentage gain in yield by increasing CO2level from 375 to
550 and 700 ppm respectively was shown by arid zone (77% and 147%) followed by
semi-arid (66%, 125%), sub-humid (22%, 35%) and humid zones (11%, 18%). This
suggests that increased CO2can increase water use efficiency. Chaudhuri et al. (1990)
and Kimball et al. (1995) have also reported that crop response to elevated CO2is rel-
atively greater when water is a limiting factor, compared to well-watered conditions.
Under full irrigation conditions the relative response of different zones is almost
similar. But if the current baseline water regimes (i.e. full irrigation in arid and semi-
arid zones and rainfed in sub-humid and humid zones) persist in future, the percent
gain in yield by increasing CO2level from 375 to 550 and 700 ppm was found to be
similar (insignificant differences) in arid, semi-arid and humid zones. But the sub-
humid zone will be most benefited (significantly higher gain in yield compared to
other zones). Reason of this can be that sub-humid zone is a rainfed zone while
arid and semi-arid zones are simulated under full irrigation. This relatively limited
water supply in sub-humid zone gets benefited from CO2enrichment. Because of
this reason when all the four areas were run under full irrigation; the differences
among zones in terms of percentage gain in yield by increase in CO2concentration
(550 and 700 ppm) became insignificant. The humid zone is also a rainfed zone but
less limited in terms of water supply (high rainfall, see Fig. 2) compared to sub-humid
zone. This is why it didn’t behave like sub-humid zone. Lawlor and Mitchell (2000)
also reported similar response in FACE (Free Air CO2Enrichment) studies, that is,
a greater effect of CO2enrichment on grain yield under drought conditions relative
to well watered conditions.
In all zones, at any specified CO2concentration, changes in the water supply from
rainfed to full irrigation resulted in yield gain, an indication that water supply is
normally a limitation to crop production in these zones.
Within a CO2level the changes in water supply from rainfed to full irrigation
shows an intense degree of responsiveness in terms of yield gain at 375 ppm CO2
level compared to 550 and 700 ppm. There was on average 324% and 195% gain in
yield in arid and semi-arid zones respectively by changing water supply from rainfed
to full irrigation in the scenario of 375 ppm CO2compared to 208% and 122% gain in
yield in the same zones at 550 ppm CO2and 150% and 87% gain in yield at 700 ppm
CO2. In sub-humid zone and humid zone this response was 31% and 18% at 375 ppm
CO2concentration; 16% and 15% at the scenario of 550 ppm CO2and 15% and 13%
at 700 ppm CO2respectively.
136 Climatic Change (2009) 94:123–142
Reduction in lenght of crop life cycle (days) from base scenario
0 1020304050
Temperature change (°C)
1
2
3
4
5
Arid Semi-arid Sub-humid Humid
Fig. 8 Change in length of crop life cycle (days) from base scenario in respective climatic zones
4.4 Length of crop life cycle in the scenarios of climate change and water
management
Temperature is an important determinant of the rate at which a plant progresses
through various phenological stages towards maturity. Rise in temperature reduced
the length of crop life cycle in all areas, though at an accelerated rate in humid
zone (Fig. 8). Husssain and Mudasser (2007) have also reported that increased
temperatures correspond to an increase in Growing Degree Days (GDDs) and a
decrease in crop life cycle. O’Brien (2000) reported the highest change in GDDs in
humid regions and the lowest in arid regions. On average there were 4, 5, 6 and 9 days
reduction in the length of crop life cycle per degree centigrade rise in temperature in
arid, semi arid, sub-humid and humid zones respectively. Reduction in crop life cycle
from the base scenario was in the order of 12, 14, 18 and 27 days in arid, semi arid,
sub-humid and humid zones respectively. In humid zone, the accelerated shrinkage
of crop life cycle was on the expense of gain in yield. Apart from gain in yield, there is
probability in the shift of cropping pattern because at prevalent temperatures wheat
crop takes more than optimal time to complete its growth and development cycle.
As temperature rises there is chance that crop completes its life cycle in less time
thus might offer the possibility of growing successive crops (moisture conditions
permitting). Rosenzweig and Liverman (1992) have also reported these outcomes
in response to temperature increase for temperate regions. This behavior can help
Fig. 9 Cumulative probability function plots of wheat yields to different sowing dates under climate
scenarios in arid and semi arid zones. a,dat 375 ppm with no change in temperature; b,eat 550 ppm
with 3◦C rise in temperature; c,fat 700 ppm with 5◦C rise in temperature in arid and semi arid zones
respectively. Bold lines represent current sowing in the respective zone
Climatic Change (2009) 94:123–142 137
(a)
Yield (tons ha-1)
3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(b)
Yield (tons ha-1)
2.5 3.0 3.5 4.0 4.5 5.0
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(c)
Yield (tons ha-1)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(d)
Yield (tons ha-1)
3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(e)
Yield (tons ha-1)
2.5 3.0 3.5 4.0 4.5 5.0 5.5
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(f)
Yield (tons ha-1)
2.5 3.0 3.5 4.0 4.5 5.0 5.5
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
15-Oct 1-Nov 15-Nov
30-Nov 30-Dec
15-Dec
138 Climatic Change (2009) 94:123–142
(a)
Yield (tons ha-1)
012345
012345
012345
01234567
0123456 87
6
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(b)
Yield (tons ha-1)
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(c)
Yield (tons ha-1)
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(d)
Yield (tons ha-1)
0246810
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(e)
Yield (tons ha-1)
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
(f)
Yield (tons ha-1)
Cumulative probability
0.0
0.2
0.4
0.6
0.8
1.0
15-Oct 1-Nov
30-Nov 15-Dec 30-Dec
15-Nov
Climatic Change (2009) 94:123–142 139
Fig. 10 Cumulative probability function plots of wheat yields at different sowing dates under climate
scenarios in sub-humid and humid zones. a,dat 375 ppm with no change in temperature; b,eat
550 ppm with 3◦C rise in temperature; c,fat 700 ppm with 5◦C rise in temperature in sub-humid and
humid zones respectively. Bold lines represent current sowing in the respective zone
increase the cropping intensity. The areas of humid zone which are currently under
mono-cropping (one crop a year) because of short growing season can become
transitional cropping zone and those currently in transitional cropping zone can
become double cropping zone if moisture conditions permit.
In arid, semi-arid and sub-humid zones, the loss of yield with rise in tem-
perature is linked with the shrinkage of the length of crop life cycle. This effect might
be overcome by selection of slower-developing varieties where water is not limiting,
but such varieties would be undesirable in today’s environment (Lawlor and Mitchell
2000).
4.5 Alterations in the sowing date as an adaptation option to offset the likely
impacts of climate change
The results of Figs. 9and 10 depict that there is high probability of increased yield
in later sowing dates compared to earlier ones under climate change. This indicates
that the sowing date of wheat in arid, semi- arid, sub-humid and humid zones will
shift towards cooler months with climate change. But different zones have different
starting points of shift.
In arid zone the current dominant time of sowing wheat is mid of December, while
in semi-arid it is the end of November. As the climate changes, the yields decline
(as discussed in previous sections). In order to reduce these losses the changes in
sowing from mid December to late December in Arid zone while in Semi-arid from
late November (current practice of sowing) to earlier December will be beneficial in
order to reduce the vulnerability. Figure 9revealed that the probability lines of later
sowings distinctly twitter out from current practice of sowing for higher yield with
climate change.
In sub-humid zone, with the climate change, the probability line of sowing date
of mid November (dominant current practice) distinctly moves towards lower yields
while the sowing in late November and in December protruded out for higher yields
(Fig. 10a, b, c). In humid zone, in the baseline climate, the current practice of sowing
(1-Nov), lies behind the sowing of 15-Oct (Fig. 10d) in terms of higher yields. As the
temperature increases as a result of climate change (Fig. 10e), the sowing of 15-Nov
protrudes out for higher yield compared to earlier sowing. As climate changes further
(Fig. 10f), the probability lines of later sowings protrude out just behind the sowing
of 15-Nov, such that the probability line of 30-Nov sowing overlaps the sowing of
15-Nov.
These results support the conclusion that the shift in the sowing date towards
cooler months can offer an opportunity to offset the likely impacts of climate change.
With each advance in sowing date the growing season length decreases compared to
the previous sowing (results not shown), in baseline scenario as well in the scenario
of climate change. This suggests that the alteration in sowing date might not interfere
with other crops grown during remainder of the year, but might offer an opportunity
of some additional time that can be utilized for land preparation or to grow an
140 Climatic Change (2009) 94:123–142
additional crop. Apart from this, cultivars that can take benefit of shorter durations
will be beneficial. Because of high economic concerns associated with adaptation
option, it will be easier to opt the alteration in sowing date as adaptation option as it
will most probably be a no cost decision, and the following and previous crops may
also not be affected.
5 Summary and conclusion
Higher temperatures have drastically reduced yields in arid, semi-arid and sub-
humid zones. But in humid zone, beneficial effects are likely to ensue with higher
temperatures up to 4◦C. This positive yield trend in humid zone can be a significant
research venture in view of the observed climatic change of temperature increase.
Increased concentration of CO2from the baseline level of 375 to 550 and 700 ppm
will have substantial positive effect on wheat yield in all climatic zones. This positive
impact will compensate the negative impact arising due to increased temperatures
of 3–4◦C in arid, semi-arid and sub-humid zones. But in humid zone at the given
CO2concentration level of 375, 550 or 700 ppm, a 4◦C increase in temperature is the
optimal one where maximum yield could be obtained.
In the scenario of baseline water regimes and baseline CO2concentration, rising
temperatures will render the sub-humid zone the most vulnerable zone and the
humid zone will be least vulnerable when compared with their respective base yields.
If all zones have homogeneous water regime (either rainfed or irrigated) at baseline
CO2concentration then the three zones (arid, semi-arid, and sub-humid) will be
equally vulnerable in yield loss. Only humid zone will respond positively. In all zones,
changes in the water supply from rainfed to full irrigation are expected to respond
positively in yield gain.
Crop response to elevated CO2will be relatively greater when water is taken as
a limiting factor, compared to well-watered conditions. This suggests that increased
CO2can increase water use efficiency.
Rise in temperature has reduced the length of crop life cycle in all areas, though
at an accelerated rate in humid zone. This reduction at an accelerated rate can help
humid zone to increase the cropping intensity, if moisture conditions will permit.
Evaluation of sowing date as an adaptation option depicts that there is high
probability of increased yield in later sowing dates compared to earlier ones under
climate change. This implies that the sowing date of wheat in arid, semi-arid, sub-
humid and humid zones will shift towards cooler months with climate change.
Acknowledgements Authors would like to acknowledge Dr. Abdul Hameed Khan, Director,
Soil Survey of Pakistan, Peshawar for his guidance in the collection of relevant soil information and
Mr. M. Asim, Scientific Officer, Wheat Programme, NARC for his help in the provision of data.
References
Aggarwal PK, Kalra N (1994) Analyzing the limitation set by climatic factors, genotype, and water
and nitrogen availability on productivity of wheat. II. Climatically potential yields and manage-
ment strategies. Field Crops Res 36:161–166
Climatic Change (2009) 94:123–142 141
Bender J, Herstein U, Black C (1999) Growth and yield responses of spring wheat to increasing CO2,
ozone and physiological stresses: a statistical analysis of ‘ESPACE-wheat’ results. Eur J Agron
10:185–195
CGIAR (2004–2005) CGIAR & climate change. http://www.cgiar.org
Chaudhuri UN, Kirkam MB, Kanemasu ET (1990) Root growth of winter wheat under elevated
carbon dioxide and drought. Crop Sci 30:853–857
Donatelli M, Bellocchi G (2001) Estimate of daily global solar radiation: new developments in the
software RadEst 3. In: Proceedings of the 2nd international symposium modelling cropping
systems, Florence, Italy, pp 213–214
Donatelli M, Bellocchi G, Fontana F (2003) RadEst 3: software to estimate daily radiation data from
commonly available meteorological variables. Eur J Agron 18:363–367
Egli DB, Bruening W (1992) Planting date and soybean yield: evaluation of environmental effect
with a crop simulation model: SOYGRO. Agric For Meteorol 62:19–29
Framji KK, Garg BC, Luthra SDL (1982) Irrigation and drainage in the world—a global review.
http://www.ciid-ciid.org-i_d_pakistan.pdf. Cited 1 June 2007
GoP (1983–1984 to 1994–1995) Crops area production (by districts). Ministry of Food, Agriculture
and Livestock (Economic Wing), Government of Pakistan, Islamabad
Hashmi AA, Shafiullah (2003) NASSD background paper: agriculture and food security. IUCN
Pakistan, Northern Areas Progamme, Gilgit
Hoogenboom G, White JW, Messina CD (2004) From genome to crop: integration through simula-
tion modeling. Field Crops Res 90:145–163
Hoogenboom G, Jones JW, Wilkens PW, Batchelor WD, Bowen WT, Hunt LA, Pickering NB, Singh
U, Godwin DC, Baer B, Boote KJ, Ritchie JT, White JW (1994) Crop models. In: Tsuji GY,
Uehara G, Balas S (eds) DSSAT version 3, vol 2. Univ. of Hawaii, Honolulu, pp 95–244
Husssain SS, Mudasser M (2007) Prospects for wheat production under changing climate in mountain
areas of Pakistan—an econometric analysis. Agric Syst 94:494–501
IPCC (2001) Climate change 2001: working group II: impacts, adaptation and vulnerability.
Cambridge University Press, New York, USA
Iqbal MM, Hussain SS, Goheer MA, Sultana H, Salik KM, Mudasser M, Khan AM (2008) Climate
change and wheat production in Pakistan: calibration, validation and application of CERES-
Wheat model. GCISC Research Report 14, Global Change Impact Studies Centre, Islamabad,
Pakistan
Johnson RC, Kanemasu ET (1983) Yield and development of winter wheat at elevated temperatures.
Agron J 75:561–565
Jones JW, Hoogenboom G, Porter CH, Boote KJ, Batchelor WD, Hunt LA, Wilkens PW, Singh U,
Gijsman AJ, Ritchie JT (2003) The DSSAT cropping system model. Eur J Agron 18:235–265
Kalra N, Aggarwal PK, Chander S, Pathak H, Choudhary R, Chaudhary A, Sehgal M, Rai HK,
Soni UA, Sharma A, Jolly M, Singh UK, Ahmed O, Hussain MZ (2003) Impacts of climate on
agriculture sector. In: Shukla PR, Sharma SK, Ravindranath NH, Garg A, Bhattacharya S (eds)
Climate change and India: vulnerability assessment and adaptation. Universities Press Pvt. Ltd.,
Hyderabad, India
Kimball BA, Pinter PJJ, Garcia RL, LaMorte RL, Wall GW, Hunsaker DJ, Wechsung G, Wechsung
F, Kartschall T (1995) Productivity and water use of wheat under free-air CO2enrichment.
Global Change Biol 1:429–442
Lawlor DW, Mitchell RAC (2000) Crop ecosystem responses to climatic change: wheat. In: Reddy
KR, Hodges HF (eds) Climate change and global crop productivity. CAB International, UK
Meinke H, Hammer GL (1995) A peanut simulation model. II. Assessing regional production
potential. Agron J 87:1093–1099
Meinke H, Hammer GL, Chapman SC (1993) A sunflower simulation model. II. Simulating produc-
tion risks in a variable subtropical environment. Agron J 85:735–742
Mustafa SZ, Yasmin S, Kisana NS (2001) Results of the national uniform yield trials 2000–2001,
pp 1–138
Mustafa SZ, Yasmin S, Kisana NS, Majid A (2000) Results of the national uniform yield trials 1999–
2000, pp 1–150
Mustafa SZ, Ahmed I, Yasmin S, Majid A (1998) Results of the national uniform yield trials 1997–98,
pp 1–137
O’Brien K (ed) (2000) Developing strategies for climate change: the UNEP country studies on
climate change impacts and adaptations assessment, report 2000–02. CICERO, Oslo, Norway
Pakistan Country Report (1994) Climate alert, vol 7, no 4, July–August 1994. http://www.climate.
org/pubs/climate_alert/articles/7.4/pakistan.shtml. Cited 1 June 2007
142 Climatic Change (2009) 94:123–142
Pannkuk CD, Stöckle CO, Papendick RI (1998) Evaluating CropSyst simulations of wheat manage-
ment in a wheat–fallow region of the US Pacific Northwest. Agric Syst 57(2):121–134
Parry ML (1978) Climatic change, agriculture and settlements. Dawson Folkestone, UK
Parry ML, Carter TR, Konjin NT (eds) (1988) The impact of climatic variations on agriculture.
Assessments in semi-arid regions, vol 2. Kluwer Academic, New York
Quadir DA, Hussain MA, Hossain MA (2004) The status of climate variation in Pakistan and its
impacts. SAARC Meteorological Research Centre (SMRC), Dhaka, Bangladesh
Qureshi A, Iglesias A (1994) Implications of global climate change for agriculture in Pakistan:
impacts on simulated wheat production. In: Rosenzweig C, Iglesias A (eds) Implications of
climate change for international agriculture: crop modeling study, EPA 230-B-94-003. U.S.
Environmental Protection Agency, Washington, DC, USA
Ritchie JT, Singh U, Godwin DC, Bowen WT (1998) Cereal growth, development and yield. In: Tsuji
GY, Hoogenboom G, and Thornton PK (eds) Understanding options for agricultural production.
Kluwer Academic, Dordrecht, The Netherlands, pp 79–97
Rosenzweig C (1985) Potential CO2-induced climate effects on North American wheat-producing
regions. Clim Change 7:367–389. doi:10.1007/BF00139053
Rosenzweig C, Liverman D (1992) Predicted effects of climate change on agriculture: a comparison
of temperate and tropical regions. In: Majumdar SK (ed) Global climate change: implications,
challenges, and mitigation measures. The Pennsylvania Academy of Sciences, Philadelphia, PA,
pp 342–361
Rounsevell MDA, Evans SP, Bullick P (1999) Climate change and agricultural soils impacts and
adaptation. In: IPCC (ed) Climate change 2001: impacts, adaptation and vulnerability, vol 43.
Cambridge University Press, Cambridge, USA, pp 683–709
Sultana H (2003) Photothermal quotient as an index to identify optimum sowing dates for different
wheat cultivars. Dissertation, University of Arid Agriculture, Rawalpindi, Pakistan
Sultana H, Mudasser M, Goheer AR, Hussain SS (2005) Agricultural potential in Karakoram region
under changing climate. In: Abstracts of international karakoram conference, organized by
Pakistan academy of geological sciences and EV-K2-CNR committee of Italy at Islmabad, 26–28
April 2005
Thompson LM (1975) Weather variability, climate change and food production. Science 188:534–541
Thornthwaite CW (1931) The climate of North America according to a new classification. Geogr Rev
21(4):633–655
Tsuji GY, Hoogenboom G, Thornton PK (eds) (1998) Understanding options for agricultural
production. Systems approaches for sustainable agricultural development. Kluwer Academic,
Dordrecht, The Netherlands. ISBN 07923-4833-8, 400 pp
Tsuji GY, Uehara G, Balas S (eds) (1994) DSSAT version 3. A decision support system for agrotech-
nology transfer, vols 1–3. Univ. of Hawaii, Honolulu
Ullah MK, Habib Z, Muhammad S (2001) Spatial distribution of reference and potential evapotran-
spiration across the Indus basin irrigation systems. International Water Management Institute,
Lahore, Pakistan (IWMI working paper 24)
Willmott CJ (1982) Some comments on the evaluation of model performance. Bull Am Meteorol
Soc 63:1309–1313
World Meteorological Organization (1979) Proceedings of the world climate conference. WMO,
Geneva
