«DIPLOMA THESIS Linking Climate Change with Food Security in the highlands of Khyber Pakhtunkhwa, Northwest Pakistan Presented by: Martin Kienzler ...»
Going further north towards the upper parts of the valley the inﬂuence of summer monsoon decreases constantly while winter and spring precipitation from western disturbances signiﬁcantly increase. Because of the sequence of high mountain ranges acting as a barrier against monsoonal moist airmasses intruding from the South, the upper Kaghan shows an increasing continentality of climate. Summer rainfall only contributes around 10 to 12%, while the proportion of winter rainfall counts up to around 45%. In Battakundi (2,670 masl) even two arid months are usually recorded in summer. Annual precipitation amounts vary between 1000 and 1200 mm. Mean annual temperatures at the valley ﬂoor of upper Kaghan only reach between 7 and 10◦ C. High winter precipitation amounts combined with cold temperatures lead to a thick snow cover, which usually leaves the upper part of Kaghan valley inaccessible and hence unpopulated from November to March. (Schickhoff, 1993, 1995) Figure 3.5 – Climate diagrams of Balakot, Naran und Battakundi. Source: Schickhoff (1993)
Vegetation Similar to the macroclimatic gradient along the valley there are diﬀerent characteristics of vegetation cover, too. In the lower part of Kaghan valley subtropical temperate sclerophyllous and pine forests dominate the natural vegetation, although they have been heavily diminished due to human deforestation. Higher elevations are characterized by moist temperate coniferous forests, mixed with deciduous trees. In the supalpine zone pencil cedar (Juniperus) woodlands occur on sunny slopes, while shady slopes are dominated by birch-forests and willow-krummholz. Alpine dwarf scrub heaths and extensive meadows are found above treeline. Due to shrinking summer precipitation the upper parts of Kaghan valley are marked by wormwood (Artemisia) steppes and Juniperus groves. Moist temperate coniferous and birch forests are restricted to shady areas.
(Schickhoff, 1995) Figure 3.6 – Climate diagrams of Balakot (left) and Chitral (right). Data obtained from PMD (2011) 3.4.2 Chitral district With an area of 14,833 km2 Chitral is the biggest and northernmost district of Khyber Pakhtunkhwa. The valley is completely isolated from the rest of Pakistan by the high mountain ranges of Hindu Kush and Hindu Raj. Only the two mountain passes Lowari (3,120 m) and Shandur La (3,738 m) connect Chitral with the other parts of Khyber Pakhtunkhwa and Gilgit-Baltistan, respectively. The district is dominated by the 350 km long valley of the Chitral River and its almost 30 side valleys. The elevation range is from 1,094 masl at Arandu to 7,708 masl at the summit of Tirich Mir, the highest mountain of the Hindu Kush range. High mountain topography is the main feature of the distrct. Around 34% of the area are situated above 4,500 m and more than 100 peaks reach elevations of above 6,000 m, (IUCN, 2001) glaciers cover 686 km2 (Kamp et al., 2004). Settlements and agriculture activities are concentrated on the narrow valley ﬂoors of the Chitral river system and are mostly found on big alluvial fans. Typically the valley width where agriculture is not hampered by slope is maximum two kilometres. The area
of agricultural activity is less than 4% of the district area. (Mulk, 1992) Climate In contrast to Kaghan valley Chitral is not reached by the Indian summer monsoon at all due to the high mountain barriers, except the extreme southern part of the district (around Drosh), where according to Mulk (1992) monsoonal rains count for about 11% of the annual rainfall. This means that almost all of the annual precipitation amount is falling during the winter and spring months (max 80-100 mm per month), while summers are almost completely dry (10 mm per month), which implements the need for irrigation. The annual amount of precipitation is around 400 mm in central Chitral (Chitral, 430 mm), reaching maximum values of 650 mm in the southern mountain ranges (Drosh, 618 mm) and leading down to only around 200 mm in the North, falling mostly as snow.
The climate of Chitral can be classiﬁed as dry semi arid, during kharif (summer) season even as arid and is distinctly continental (Chaudhry & Rasul, 2004). Indeed precipitation amounts rise with elevation: At altitudes of above 4,500 m precipitation amounts are estimated to be more than 2,000 mm per year. The snowline rises northwards from around 4,800 to 5,300 m (Kamp et al., 2004).
Mean annual temperatures lie around 16◦ C at the valley ﬂoor. Maximum summer temperatures usually reach 35-36◦ C, while July is the hottest month. Mean minimum temperatures in January drop down to -1◦ C. (PMD, 2011) Vegetation In Chitral district forest cover accounts for 4% of the total area and forest areas are concentrated to southern Chitral. Dry oak woodlands cover around one third of the forest area and occur at elevations between 1,400 and 1,680 m. Dry temperate coniferous forests (deodar, pine, ﬁr, spruce) are found at higher elevations (1,760 - 3,500 m). Above 3,500 m sub-alpine birch and willow shrubs are encountered and spread up to 3,800 m.
At higher altitudes there are only moist alpine pastures. In upper Chitral, where landscape is extremely denuded and conditions are semi-arid, degraded broad-leave forests are found at favourable locations. Else only scattered remnants of Juniper forests are found. (IUCN, 2001; Nadeem et al., 2009) Agriculture and irrigation Because of the relatively low precipitation amounts especially during the summer months, agriculture in Chitral district is not possible without irrigation. In 2000 only 1.6% of the cropped area was under rainfed conditions and wheat yields of rainfed agriculture were only one ﬁfth of the yields gained under irrigation. Thus almost all of the around 27,000 ha of cultivated area receives water of one of more than 1,000 irrigation channels, of which around three fourth were built by the local population. The length of the channels is between two and eight kilometres with an average of four kilometres, each of them irrigating between 10 to 70 hectares. Institutions like the Aga Khan Rural Support
Chapter 3. The province of Khyber Pakhtunkhwa - Study Area
Program (AKRSP) have begun to construct irrigation channels on a larger scale and improving or extending the existing ones. (Mulk, 1992; IUCN, 2001) Double cropping is possible at about half of the cultivated area in the lower regions.
Subsistence agriculture of Chitral is dominated by cereals, which are cultivated on about 82% of the cropping area. Nevertheless about 30,000 tons of wheat have to be imported every year to satisfy the local demand. Main crops are wheat, maize, barley and rice, which are grown on 40,6%, 28,5%, 18,7% and 12,2% respectively of the area. Compared to other parts of Khyber Pakhtunkhwa, yields per hectare of Chitral are favourable, for example it has the highest maize yields (2,790 kg/ha) of all the districts.
Another very important sector of agriculture is livestock, though only for subsistencial use. Livestock are held in small herds of about 4 to 10 animals per household. For this reason, a considerable amount of the crops is grown for fodder purposes. (IUCN, 2001) 4 Climate data
4.1 Observation data - CRU (Climatic Research Unit) The data which is used in this study originate from the database of the Tyndall Centre for Climate Change Research of the University of East Anglia in Norwich, United Kingdom.
The Climate Research Unit (CRU) oﬀers several updates of this database, here the most recent of them is used, namely CRU TS 3.1, which is an update of the CRU TS 2.1 dataset presented by Mitchell & Jones (2005).
The CRU data extends over the period 1901-2009 and represents a 0.5◦ lat-lon gridded dataset of monthly surface climate, which covers all global land areas, excluding Antarctica. Hence it forms one of the most extensive and detailed climate observation datasets available. It is based on the data of a global network of meteorological observing stations and on statistical interpolation methods.
New et al. (2000) constructed this database with the help of an anomaly approach, in the ﬁrst instance to achieve longer time series while going back in time compared to station data. For this purpose ﬁrst of all monthly anomalies were calculated relative to the climatology of the 1961 to 1990 standard normal period calculated by New et al.
(1999). These anomaly grids were then combined with a “high-resolution mean monthly climatology” (New et al., 2000) in order to get monthly surface climatic values as a result. Strict temporal ﬁdelity is ensured by this. Monthly anomalies are rather dependent of large-scale circulation patterns than of physiographic control.
Mitchell & Jones (2005) updated the described dataset through a homogenization method. For this neighbouring stations were used to “construct a reference series against which a candidate series may be compared.” The selection of the neighbouring stations occurs through a correlation method. The problem of this method is that one needs to have a set of stations, known to be homogeneous, with which to start. If new stations are incorporated into the dataset it is important to ensure that variations are caused only by variations in climate. The aim is to maximize the number of stations used for gridding.
The datasets comprise a suite of seven climate elements: precipitation, mean temperature, diurnal temperature range, wet-day frequency, vapor pressure, cloud cover and ground frost frequency. The ﬁrst three of them are primary variables and were interpolated directly from station observations.
For this study the CRU data was used for the following reasons: First the dataset has longer temporal coverage than other datasets with similar spatial resolution, covering more than 100 years. Second it has higher spatial resolution than other datasets with similar temporal range. (New et al., 2000)
Chapter 4. Climate data
Area The area which is considered for trend analysis extends from 60.5◦ East to 78.0◦ East longitudinal and from 23.5◦ North to 37.5◦ North latitudinal. The result is 1,044 grid boxes, which include all of Pakistan, the most parts of Afghanistan and small parts of India, Iran, Turkmenistan, Uzbekistan, Tajikistan and China.
For the calculation of the trend curves of Khyber Pakhtunkhwa only 55 grid boxes were used, which describe the area of Khyber Pakhtunkhwa and lie between 69◦ and 74◦ East and between 31◦ and 37◦ North. See ﬁgure 4.1 for the spatial extension.
Figure 4.1 – The grid of the regarded area for CRU TS 3.
1 trend analysis. The red dots represent the grids of Khyber Pakhtunkhwa Problems During the gridding process a correlation decay distance was deﬁned for each variable.
This distance amounts to 450 km for precipitation and 1,200 km for temperature. That means that there might be up to 900 km between actually stations for precipitation measurements and 1,200 km for temperature values respectively. Hence the calculated values for the gridboxes in between may be deﬁcient. Furthermore variations of the climate stations may be generated through environmental changes (e.g. urbanization), methodical changes, technical changes or location changes which never can be excluded.
(Mitchell & Jones, 2005) As already mentioned above the distribution of climate stations especially over the mountainous parts of Northern Pakistan is extremely sparse. For example there are several grids which do not contain a single climate station. These grids are then interpolated from neighbouring grids. This implicates that the small-scale but extreme diﬀerences in relief are not taken into account with this method, which could lead to considerable
faults. For this reason another dataset shall be used, which even has a higher resolution and includes physiographic features (see section 4.3) of the complex topography of this region.
4.2 Observation data - PMD (Pakistan Meteorological Department) station data For comparison and validation of the two gridded datasets (see section 4.1 and 4.3) the observational data of available meteorological stations is consulted. The Climate Data Processing Centre (CDPC) of the Pakistan Meteorological Department (PMD, PMD (2011)) supports 13 climate stations within the borders of Khyber Pakhtunkhwa. (see table 4.1 and ﬁgure4.2)
Table 4.1 – The PMD meteorological stations of KP.
Data obtained from PMD (2011) Nine of the thirteen meteorological stations lie at an altitude of more than 1,000 meters above sea level, one of them even at more than 2,000 masl. But anyhow there do not exist any high-altitude stations because all the stations are at the bottom of the mountain valleys. Though the distribution of the stations represents the diﬀerent climate zones, which exist within Khyber Pakhtunkhwa quite good. The time span of the observations are diﬀerent: Some stations oﬀer time series from 1931 on while others only started in the 1960s or 1970s. However gaps within the time series unfortunately are not unusual.
Figure 4.2 – The distribution of PMD meteorological stations in Khyber Pakhtunkhwa.
Source: modiﬁed from PMD (2011)
4.3 Model data - REMO (Regional Climate Model, RCM) In a remote region like mountainous Khyber Pakhtunkhwa, where meteorological stations are scarce and hence climatological observations do not exist for large areas, it often makes more sense to back on climate model simulations, too. Furthermore the resolution of the CRU data is not high enough to describe the distinct small-scale diﬀerences of climate in KP properly. And the CRU data only represents past climate.