Impact of climate change on agriculture during winter season over Pakistan ()
1. INTRODUCTION
Agriculture has always been the most important sector of Pakistan’s economy. In the agro climatic classification of Pakistan, more than two-third of Pakistan lies in semiarid to arid zones [1]. About 70% of our population is living in rural areas, where most of them along with livelihood depend on agriculture production. Approximately 50% of the total national labor force is directly engaged in agriculture [2]. Therefore, majority of the people living in arid and semi-arid areas are totally depending on agro-pastoral activities for their survival. Currently in climate change scenario Pakistan-like other developing world, is faced with the challenges of (being affected by variability of precipitation and risen temperatures) land degradation or desertification and other environmental problems like soil erosion, loss of soil fertility, flash floods, salinity, deforestation and associated loss of biodiversity and carbon sequestration [3]. Agriculture is the most vulnerable sector to climate change. Productivity of agriculture is being affected by a number of climatic factors and some indirect factors including rainfall pattern, temperature hike, changes in sowing and harvesting dates, water availability, evapo-transpiration and land suitability. All these elements have impact on crop yield and agricultural productivity [4].
Agriculture in Pakistan is dependent on rainfall as well as irrigation water. Water mainly meets from seasonal rainfall as well as melting of snow and ice from the glaciers. Pakistan has developed the world’s largest contiguous canal irrigation system. Pakistan is covered on the north by Himalaya, Karakoram and Hindukush, which host the world’s third largest snow/ice reserves. These mountains are the water tanks over the roof that provides water to the reservoirs. The environment has given the operational control of this tank in terms of temperature after the strong buildup of greenhouse gases [5].
Winter brings lot of snow over the northern mountains which melts in early summer and maintains the sustainable river flows for power generation and irrigation before the onset of the summer monsoon. In addition to solid precipitation over hilly areas, winter rain bearing systems yield substantial rainfall in sub-mountainous and low elevation plains including arid plains of Balochistan. Generally northern half gets about five times more precipitation in winter than the southern half [6]. Past analysis of precipitation data has shown a slightly decreasing trend for the northern parts of the country. On the other hand the situation for southern parts of the country is becoming better in terms of precipitation and temperature as well. But the northern belt is the major source of water to the Indus, the leading river in the country. The present increase in temperature may be major augmenting force behind the sharp decrease in this treasure of solid water. For a country which is already facing problems because of ill management in water distribution and inadequate water reservoirs, it could lead to the collapse of the local agriculture system in the time to come [6]. Higher rainfall variance seems to be the main factor behind dry-land yield fluctuations. Amount and distribution of Rainfall during crop season are important. Distribution of rainfall becomes more significant for the lands with low water holding capability and also in the seasons with adequate soil moisture available at planting [7].
Increasing temperatures may have a positive impact on agriculture in the mountain areas, for instance, through shortening of growing period for the winter season crops. Winter crops (e.g. wheat), in the high mountain areas, do not even reach to maturity in most cases and such crop is harvested premature to be used as fodder. The shortening of the growing season length due to rising temperature could be beneficial in the mountain areas as it would help the winter crops in timely maturity and as such would allow the crop to mature in the optimal period of time, with beneficial effects on crop area and yields [8].
There is high level of confidence that recent regional changes (rising tendency) in temperature have discernable impacts on precipitation, evaporation, stream flow, runoff and other elements of hydrological cycles [9]. Under increased Green House Gasses (GHG) concentrations, global climate models also exhibit enhanced intensity and shorter return periods of extreme events [10,11]. Recorded extreme events during the last decade of 20th century depict consistency in terms of intensity and frequency. The history’s worst drought with extremely high air temperatures and without snow cover during winter 2001, history’s worst flash floods in July 23, 2001 in Rawalpindi/Islamabad because of Cloud burst, are the few sound evidences of increased intensity of extreme events [12].
2. DATA AND METHODOLOGY
The basic water equation for the column is given as:
(1)
Here E represents evapotranspiration, P is precipitation, 
is the horizontal divergence operator and W is vertically integrated water content per unit area given by:
(2)
Whereas, Q is the vertically integrated moisture flux given by
(3)
Whereas q is the specific humidity in kg/kg, dp is change in pressure, ps is surface pressure, pt is pressure at top of an atmospheric column taken to be at 300 hPa where q becomes negligible, g is the acceleration due to gravity, 9.8 m/s2, and V is the horizontal wind vector defined as:
(4)
Here u and v are the eastward and northward wind components respectively. Bold variables indicate vector representation.
The National Center for Environmental Prediction and National Center for Atmospheric Research (NECPNCAR) pressure level data were obtained at 6 hours intervals, 00, 06, 12 and 18 Z. Four atmospheric variables such as the specific humidity q, zonal wind u, the meridional wind v and the surface pressure Ps are used for this study. All the variables except surface pressure are located on a 144 × 73 horizontal grid with a resolution of 2.5 degrees of latitude and longitude, and at 8 levels in the vertical, 1000, 925, 850, 700, 600, 500, 400, 300 hPa. Above 300 hPa, the air is very dry and contributes little to water vapor transport. Surface pressure is also available on 144 × 73 horizontal grids.
The vertically integrated Moisture Flux Convergence (MFC) into the region was calculated by using Green’s Divergence Theorem.
(5)
where n is the unit outward normal vector on the domain boundaries and l is the length of the line segment along the boundary. For domain average MFC, the right side is an integral performed around the entire domain. Evaluating the right hand side of Eq.5 around the chosen boundary, then dividing by the area of the entire domain will give average convergence rate of the domain. The boundary chosen is very coarse with a grid spacing 2.5˚ × 2.5˚. The total land area of the domain used in this study is 3.14 × 1012 m2. Vertically integrated MFC from surface to 300 hPa for individual grid boxes (2.5˚ × 2.5˚) have also been computed and used for examining the spatial distribution of MFC. (Figure 1)
Vertical integration around the region becomes difficult due to complex terrain and uneven surface, as surface pressure does not necessarily correspond to 1000 hPa, the lowest level pressure available in the data sets. Since some points having 1000 hPa were underground while others were in the atmosphere, qV values were linearly interpolated or extrapolated as required. The Global Precipitation Climatologically Centre (GPCC) version 5 data with grid point at 0.5˚ resolution of latitude and longitude for precipitation and NCEP reanalysis data for geo-potential height at different levels and zonal wind at 200 hPa have been used in the study. Precipitation and temperature anomalies have been computed from long term (1901-2009) GPCC precipitation average and (1949- 2009) NCEP temperature data respectively. All the analysis has been made on the winter season starting from December (of previous year) to February (of current year). This implies that precipitation of 1962 consist of average precipitation of December, 1961 to February, 1962. All the calculation monthly as well as seasonal was made on monthly basis. The calculation has been made by using the software developed by the International Research Institute, University of Columbia, which is available on line at:
http://iridl.ldeo.columbia.edu
3. THE STUDY PERIOD (1961-2009)
The causes of climate change in any region can be linked with variability of moisture in atmospheric and precipitation on surface. Both these factors are closely associated with variation of air temperature in the region. Figure 2 depicted anomaly of precipitation in the domain during study period. Precipitation anomalies can easily be distributed in the three phases such as: 1) No radical change during the period 1961-1985 named as before climate change scenario, no prominent variability in precipitation is observed during this phase except for few years; 2) The period 1986-1999, during which a large variation of precipitation is observed in the domain, it is assumed that this period is considered as transition period. This period is considered as rapid climate pattern change scenario in view of surface pressure and air temperatures of Atlantic and Pacific Oceans; and 3) Drastic increase in precipitation is observed during 2000-2009 after climate change scenario. In this phase it is observed that precipitation in the region has drastically increased with increase in the possibility of uncertainty of weather conditions. The study is based on the variation of different meteorological parameters during these different phases and it has been tried to investigate causes of these variations in precipitation. The cause of this drastic increase in precipitation in the region has been framed by com-
Figure 1. Boundary line in the map represents the area used for computing MFC in this study. Grids interval = 2.5˚ × 2.5˚ (longitude × latitude).
Figure 2. Anomalies of precipitation in the region during 1961-2009. Polynomial trend line represents with black steady line.
paring different meteorological parameters during these different three phases.
4. ANALYSIS OF DIFFERENT METEOROLOGICAL PARAMETERS
4.1. Moisture Flux Convergence (MFC)
Average seasonal (Dec-Feb) and five year running average MFC over entire domain has been computed and depicted in Figure 3. Highest moisture transported seasons were 1997-1998 and 1972-1973 with the values of 19.69 cm and 19.41 cm per month respectively. The least MFC seasons were 1999-2000 and 1996-1997 with 10.39 and 10.69 cm/month. Moisture flux convergence seasonal (Dec-Feb) time series fluctuate between 10.39 to 19.69 cm/month with an average value of 14.56 cm/ month. Highest value of five years MFC has been computed during late seventy’s and least was found in late ninety’s. However, decreasing trend in five year average MFC is prominent from early ninety’s. Figure 4 depicted monthly and seasonal (Dec-Feb) MFC of three different periods. During the season, February is considered as highest moisture transported month whereas during December it has the least values. Comparison showed that moisture transported decreased throught out the season as well as in corresponding months during the period of 2000-2009. Five year running average MFC of entire domain showed decreasing trend from early nineties. This corresponds that moisture transport into the domain is decreasing.
In December, moisture transported has decreased during after climate change scenario, and it has remained highest during transition period. In January, moisture transported during transition and before climate change scenario is almost same whereas during after climate change scenario it decreased. In February, moisture transported during before climate change scenario is highest while during after climate change scenario it became decreased. In seasonal comparison (Dec-Feb), moisture transported has been decreased after climate change scenario with 1 cm/month. In transition period (1986- 1999), average moisture transported in December is highest as compared to before and after climate change scenario because of too much variation in moisture transport during this period.
Figure 6 depicted spatial distribution of moisture flux convergence of entire domain. Positive values indicate moisture convergence and negative values indicates moisture divergence region. The moisture divergence regions referred as source of moisture and moisture convergence region referred as sink of moisture. There are two main source of moisture for Pakistan; one is from the south i.e. Arabian Sea during both periods and other from the west. The source of moisture indicates the track of approaching weather systems in the region. There are two moisture convergence regions in both before and after climate change scenarios, one is over extreme north extending towards south east up to Kashmir and other is over central Afghanistan extended towards east crossing from Central Pakistan and Western Punjab of India. Both periods showed almost same regions of moisture transport. This indicates that no remarkable variation in the region of moisture sink/source has been observed during the study period. However, slightly decrease in moisture transported region over Central Afghanistan and Central Pakistan has also been observed during after climate change period. Figure 6 depicted MFC anomalies along with three different, most variability values during before climate change scenario (A), transition period (B), and after climate change scenario(C). It is observed that most of the values lie in (A) from 1.75 to –2.0 cm, in (B) from 3.39 to –2.3 cm and in (C) from 0 to –2 cm. It is clear evidence that during after climate change scenario, the frequency of more transposed moisture is increased with reduced moisture transported in the entire domain. The average moisture transported is 0.78 cm/month which is about 0.81 cm/month is higher than before climate