«DIPLOMA THESIS Linking Climate Change with Food Security in the highlands of Khyber Pakhtunkhwa, Northwest Pakistan Presented by: Martin Kienzler ...»
Interviews were conducted during a ﬁeld trip to Khyber Pakhtunkhwa and in both of the case studies described in chapter 3.4. The reasons why these case studies were selected already have been explained. Within the areas of the case studies several places were selected for questioning. The places should represent diﬀerent altitudes to ﬁnd out how agriculture would be aﬀected by climate change in diﬀerent elevations of the mountain region. Accessibility however was the limiting factor of visiting these places.
They are listed in table 5.1.
For Kaghan valley interviews were arranged in two diﬀerent villages: Kaghan valley is situated at the bottom of the lower valley at an altitude of approximately 2,000 masl.
Battakundi is a small village in the upper part of the valley and lies at around 2,700 m altitude. Just above the village the Lalazar Plateau is found at around 3,200 masl, one of the most elevated places of the region, where agriculture is practised.
In Chitral district, where cropping generally is more variable than around Kaghan, several villages were selected for qualitative research. Mastuj is a village in the northeastern part of the valley near the border to the Northern Areas or Gilgit-Baltistan. It is situated at an altitude of around 2,400 m. Within the valley of Garam Chashma, which spreads out from Chitral town to the West two villages were selected: Rugi (2,100 masl)
Chapter 5. Methodological approach
and Shoghor (1,900 masl). Further South interviews were performed inside the Kalash valleys (Rumbur, 1,900 masl and Bumburet, 1,960 masl), which are side valleys of the main valley to the West, and in Ayun (1,400 masl), a village at the place where the Kalash valleys lead into the Chitral valley.
This forms a basis of samples at an elevation range of between 1,400 and 2,700 masl.
In each village three to ﬁve persons were interviewed, either at public places like the bazaar, at their homes or within their ﬁelds.
The interviews were preferentially conducted on the elder people of the villages to ensure a certain time span (at the best case more than 30 years). Because of traditional and cultural issues only male farmers could be interviewed. The questionnaire (see Appendix A) is somehow structured, but the interviews did not follow always exactly the same question. They rather were adapted to the respective course of the discussion. The questions were asked after the following structure. The ﬁrst part of the questionnaire deals with cropping characteristics in general. It aims to ﬁnd out, if there are any changes in variables like length and onset of sowing and harvesting seasons respectively or if the patterns of cropping did change in any respect. The answers ideally should indirectly advert to climatic changes. The second part is concerning livestock, in most regions the second pillar for food security. The third part ﬁnally focusses on direct questions on climate issues. It aims to detect the perception of the key informants to changes in day and night temperatures, precipitation, extreme events and seasonal diﬀerences. The climate issue is composed on the end of the questionnaire and instead at the beginning questions were asked concerning issues indirectly related to climate. This is to avoid the “possibility that respondents would give answers ’preferred’ by the researchers” (Thomas et al., 2007).
Table 5.1 – The places, where qualitative interviews were conducted 6 Results The results are illustrated by the following plots.
All the plots are generated with the help of Generic Mapping Tool, a software developed for visualizing maps.
This chapter is arranged as follows: For each of the two climate variables (surface temperature and precipitation) the results are presented separately. Mean values, annual trends and seasonal trends are described successively in both cases. By this way an overview of the whole region as well as a focus to the two case studies is shown. The next step is the presentation of future climate prediction. The last section of the chapter deals with the results of the questionnaire.
6.1 Temperature 6.1.1 Mean climatology The distribution of annual mean temperatures of Pakistan and the adjoining areas (Afghanistan, Iran, India, China) strictly follows the relief characteristics and reveals extreme diﬀerences in climate between the diﬀerent parts of the country. The vast plains along the Indus river in the Southeast of the country are marked by very high annual temperatures, not least because of the position of the summer heat low and the existence of the desert Thar. Mean annual temperatures are well above 25◦ C, REMO even simulates annual temperatures of more than 30◦ C in some parts. (see ﬁgure 6.1).
Temperatures decrease north- and northwestwards in direction of the relief gradient.
The mountain regions along the Afghan border in western Pakistan, the foothills of Himalaya and Hindu Kush in the North and the valleys between these high mountain ranges show annual mean temperatures of between 9 and 18◦ C, mainly depending on elevation. Mean temperatures of the high mountain areas hardly reach positive values.
Here REMO simulates far lower temperatures than the CRU data shows.
It is obvious that the REMO model output distinctly includes small-scale relief diﬀerences. So the small and narrow valleys between the high mountain ranges of the Hindu Kush - Karakoram - Himalaya system are not taken into account by the CRU data and therefore are considered too cold. Generally the two gridded datasets are according with each other quite well. Only the REMO model somehow underestimates the minimum temperatures in high mountain regions and on the other side overestimates temperatures in desert areas.
For regarding seasonal diﬀerences in mean temperature the REMO model output was consulted because of the higher resolution. Seasonal values accentuate the distinct temperature gradient between the North and the South (Southeast) of the country even more (see ﬁgure 6.2). During the summer months in most parts of the Indus plains and
Figure 6.1 – Annual mean temperature over Pakistan.
Left: CRU data. Right: REMO model output adjoining Ganges plains and even in the Southwestern edge of the country, the border triangle Pakistan - Afghanistan - Iran, temperatures stay well above 33◦ C. This is the time when the heat low stabilizes over the desert Thar and monsoon comes to its climax.
In these regions, June is the hottest month before monsoonal rainfall starts. Maximum temperatures often exceed 50◦ C, whereby this region is one of the hottest places on earth.
At the same time temperatures within the highest mountain ranges still are below zero, while even the small mountain valleys suﬀer temperatures of up to 30◦ C. In the regions where monsoon can not unfold its inﬂuence and summers are comparably dry, July is the hottest month of the year. In autumn the plains still show mean temperatures of 25 to more than 30◦ C. The foothills of the Himalaya already get cooler when the extremely high humidity of the monsoon months recedes in September. In the mountains temperatures decrease rapidly after a short summer. When temperatures fall far below zero in the northern highlands with the beginning of winter, the DJF mean temperatures in the plains stay between 12 and 18◦ C and exceed 25◦ C again already in early spring, when the heat low begins to develop over the Indian subcontinent. The warming in the valleys of the North and the submountainous regions in the West takes place more slowly.
6.1.2 Trends The development of the climate parameters during the twentieth century is the main focus of this thesis. To generate a climatic curve for Khyber Pakhtunkhwa the values of the monthly gridded CRU TS 3.1 dataset were used. For each month the arithmetic mean of all grids which represent the area of Khyber Pakhtunkhwa was calculated after equation 5.1. After that annual means were generated. These means had to be subtrated from the 1961 to 1990 average to get anomaly values. These anomalies are shown in ﬁgure
6.3 for the period 1901 to 2009 along with an 11-year running average (red line). The
Figure 6.2 – Seasonal mean temperature over Pakistan as simulated by the REMO climate model.
MAM = spring, JJA = summer, SON = autumn and DJF = winter.
values are calculated out of 55 grids, whose position is shown in ﬁgure 4.1.
The course of Khyber Pakhtunkhwa’s mean temperature shows a very high interannual variability. Though some similarities with the development of the global mean temperature (see ﬁgure 2.1) can be recognized. The century starts with a decade of rising temperatures and a ﬁrst secondary peak around 1915. Thereafter temperatures decrease sharply to the coldest period of the twentieth century in the early 1920s. The next twenty years are marked by a distinct warming and the temperature maximum of the twentieth century is reached in the ﬁrst half of the forties. This observation is valid for most parts of the globe and is known as “mid-century warming” (Latif, 2009). A possible reason for this is an intern variation of climate like the multidecadal variability of the Northern Atlantic Ocean. Similar to global mean temperatures there is a distinct cooling of surface temperatures between 1945 and 1970, before temperatures rise again for some years. A reason for this period of decreasing temperatures could be the augmented
Chapter 6. Results
emissions of aerosols which causes the “global dimming” or “ aerosol eﬀect” (Latif, 2009).
After this the global mean temperature was rising again from the early seventies on until now. In contrast the mean temperature of Khyber Pakhtunkhwa stayed more or less at a constant level until the ﬁrst half of the 1990s after having reached another secondary peak at around 1975. But nevertheless it shows a strong increase in the last 15 years, which is again comparable to the development of global temperatures in the last decade.
Figure 6.3 – Mean annual temperature averaged over Khyber Pakhtunkhwa, 1901-2009 anomalies of the 1961-1990 mean (◦ K).
Red line = 11-year running average.
Data obtained from Jones & Harris (2008) Figure 6.4 shows the linear trend for diﬀerent periods of the twentieth century and the ﬁrst decade of the twenty-ﬁrst century respectively for the CRU TS 3.1 dataset. The trend values were extrapolated to a 100 year period to enhance comparison of periods with a diﬀerent time span. The black dots represent the grids, in which the temperature trends are signiﬁcant at the 95%-level. Regarding the trend over the whole period 1901 to 2009, it becomes obvious that there is no constant warming trend all over the region.
Most parts of the country are aﬀected by an increase of temperatures of up to 2◦ C per century, the strongest warming being in the southwestern part of Pakistan. In contrast two regions show rather negative temperature trends: The region of the Sahed Koh range around Parachinar at the western border and a small region at the foothills of the Himalaya at the border to India.
To detect diﬀerent developments during the last 109 years the whole period was split into the periods 1901 to 1940, 1941 to 1970, 1971 to 2000 and 1981 to 2009. The century started with a period of quite diﬀerent trends: The eastern part of the country suﬀered
Figure 6.4 – Temperature trend over Pakistan for diﬀerent periods extrapolated to ◦ C per 100 years.
The black dots represent grids, which are signiﬁcant to 95% (α = 0.95) a distinct warming of +1 to +2◦ C with warming hot spots over the southeastern Indus plains and the northeastern mountains. At the same time the western half of Pakistan experienced a strong cooling trend: the regions along the Afghan border show negative trends of up to -2◦ C per century between 1901 and 1940.
The period 1941 to 1970 globally marks an era when temperatures generally tended to drop down instead of going up. The same is valid for most parts of Pakistan: Especially
Chapter 6. Results
in the central part temperatures decreased at a rate of -2 to -3◦ C, in some parts even up to
-4◦ C per 100 years. Only in the Southwest near the Iranian border and in the Karakoram ranges in the farthest North temperatures went up during these years. During the next thirty years, between 1971 and 2000, a distinct warming took place again. The central parts of the country suﬀered a moderate temperature rise of less than +2◦ C per century.
The hot spot of warming seems to be found at the southern part of the border area between Pakistan and Afghanistan, where increasing rates of more than +6◦ C per 100 years are recorded. The Northern part and most of Khyber Pakhtunkhwa show warming rates of less than +1◦ C. There are two regions where in contrast temperatures seem to fall dramatically: In a small area near the estuary of the Indus river into the Arabian Sea, which indeed is not signiﬁcant at the 95%-level and a larger area around Parachinar at the Afghan border which instead is signiﬁcant. Here temperatures drop at a rate of up to
-5◦ C per century, which is rather unusual. This draws the picture of a stark contrast on a comparable small scale which has to be discussed later (see chapter 7). Regarding the most actual period 1981 to 2009 a similar picture arises. The only diﬀerence is that the cooling rates in these mentioned regions are not as high as during the previous period.
But the warming throughout the country is even stronger than before. Especially in the northern mountainous parts temperatures rise as fast as never before. Maximum increasing rates are recorded in the most western part of the country and add up to more than +7◦ C per century.
The conclusion which can be drawn from these results is that generally after a warming period during the ﬁrst forty years of the twentieth century temperatures were going down for about thirty years to rise again at an distinctly accelerated rate towards the end of the century and beginning of the twenty-ﬁrst century respectively.