The Oxford Research Encyclopedia of Natural Hazard Science will be available via subscription on April 26. Visit About to learn more, meet the editorial board, learn about subscriber services.

Dismiss
Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, NATURAL HAZARD SCIENCE (oxfordre.com/naturalhazardscience). (c) Oxford University Press USA, 2019. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 20 April 2019

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush Region

Summary and Keywords

This chapter analyzes the impacts of climate change on flood factors and extent of associated damages in the Hindu Kush (HK) region. HK mountains system is located in the west of the Himalayas and Karakorum. It is the greatest watershed of the River Kabul, River Chitral, River Panjkora, and River Swat in the eastern Hindu Kush and River Amu in western Hindu Kush. The Hindu Kush system hosts numerous glaciers, snow-clad mountains, and fertile river valleys; it also supports large populations and provides year-round water to recharge streams and rivers. The study region is vulnerable to a wide range of hazards including floods, earthquakes, landslides, desertification, and drought. Flash floods and riverine floods are the deadliest extreme hydro-meteorological events. The upper reaches experience characteristics of flash flooding, whereas the lower reach is where river floods occur. Flash floods are more destructive and sudden. Almost every year in summer, monsoonal rainfall and high temperature join hands with heavy melting of glaciers and snow accelerating discharge in the river system. In the face of climate change, a significant correlation between rainfall patterns, trends in temperature, and resultant peaks in river discharge have been recorded. A rising trend was found in temperature, which leads to early and rapid melting of glaciers and snow in the headwater region. The analysis reveals that during the past three decades, radical changes in the behavior of numerous valley glaciers have been noted. In addition, the spatial and temporal scales of violent weather events have been growing, since the 1980s. Such changes in water regimes including the frequent but substantial increase in heavy precipitation events and rapid melting of snow in the headwater region, siltation in active channels, excessive deforestation, and human encroachments onto the active flood channel have further escalated the flooding events. The HK region is beyond the reach of existing weather RADAR network, and hence forecasting and early warning is ineffective. Here, almost every year, the floods cause damages to infrastructure, scarce farmland, and sources of livelihood.

Keywords: flood, physical factors, human-induced factors, Hindu Kush, damages, climate change

Introduction

Globally, flooding is a serious hydro-meteorological phenomenon (Rahman & Khan, 2011). In terms of losses and extent, flooding is considered to be the most devastating disaster (Ali, 2007; Zhang et al., 2011). In the face of a changing climate, the frequency and magnitude of floods have increased, since the 1980s (Rahman & Khan, 2013). The ongoing changing climate can lead to frequent disastrous floods, and the increasing anthropogenic activities in the river proximity and meagre flood risk reduction strategies have further intensified flooding events (Rahman & Khan, 2013). It is evident from the scientific literature that the disastrous floods may become more widespread, with the increasing extreme precipitation events as a result of climate change (IPCC, 2014). When incidents of floods occur, water overflows the levees and inundates the active floodplain (Hunter et al., 2005; Lehner et al., 2006) and cause damages to people and their properties (Nathan, 2008). In order to minimize the adverse impacts of floods, non-structural and structural measures are implemented (Montz & Gruntfest, 2002) as flood risk reduction strategies (Mazzorana, Hubl, & Fuchs, 2009).

In 2010 Pakistan was hit by a devastating flood event (Syvitski & Brakenridge, 2013; Mahmood, Khan, & Mayo, 2016a), where over 1,900 people lost their lives. Floods inundated 21% of the country, and 14 million people were faced with billions of dollars in losses (Tariq & Giesen, 2012). The nature and direct impact of the flooding varied from place to place depending on the spatial variation in land-use land cover, physiography, and population density (Mahmood, Khan, & Ullah, 2016b). The primary cause of this flood was the four-day (July 27 to 30) wet spell along with melting of glaciers and snow especially in the Hindu Kush region.

Hindu Kush is a region of high mountain systems, located to the west of Himalayas and Karakorum (Figure 1; Rahman & Khan, 2013). Due to its dignity as great mountain system, it has been called a “roof of the world.” It is separated from the Himalaya at the Pamir knot, where the border of Pakistan, Afghanistan, Tajikistan, and China meet (Khan, 2003). It stretches from northeast to southwest to 1,600 km in length and 323 km in width (Shams, 2006). At the north, it separates Pakistan from Tajikistan through a narrow strip called “Wakhan corridor” (Qamer et al., 2012). Tirich Mir is the highest peak having elevation of 7,700 masl (Rahman, 2010; Figure 2). It is the greatest watershed of the River Kabul, River Chitral, River Panjkora, and River Swat in the eastern Hindu Kush and River Amu in western Hindu Kush. The Hindu Kush system hosts numerous glaciers, snow-clad mountains, and fertile river valleys; this system supports large populations and provides year-round water to replenish streams and rivers. This studied region is vulnerable to a wide range of hazards including floods, earthquakes, landslides, desertification, and drought. In the HK region, flash and riverine floods are deadliest extreme hydro-meteorological events. The upper reaches experience characteristics of flash floods, whereas the lower reaches have river floods. Here, flash floods are more destructive and sudden.

Climatically, the HK region has three climatic zones: humid zone, sub-humid zone, and a zone of semi-arid as one moves from higher elevation in the north toward the south. In the HK region, the average annual precipitation varies from over 2,000 mm in the humid to sub-humid areas to less than 200 mm in the semi-arid areas. Similarly, there is also a wide variation in the seasonal temperature. During winter, mean low temperatures fall below –2°C at Dir, Kalam, and Malam Jabba met stations, whereas in summer the maximum temperature exceeds 42°C at Saidu and Timergara met station. Extreme winter temperature provides a foundation for the snow and glacier accumulation, while high summer temperature accelerates melting of snow and glaciers. It is therefore such seasonal variation in rainfall and snow that poses a variety of disaster risks, out of which flooding is the most frequent and widespread (Gupta & Sah, 2008; Qamer et al., 2012).

Geomorphologically, gradient of the river channel is steep in the upper reaches, while comparatively gentle in the lower reaches of Hindu Kush. Due to physical terrain, flash floods dominate the upstream areas and river floods in the gently sloping low-lying areas. In the HK region, fluctuation in river discharge is not only attributed to heavy rainfall (Ramos & Reis, 2002) but also to geophysical, hydro-meteorological, and human-induced factors (Zhang et al., 2006; Rahman & Khan, 2011). Therefore, during 1980 to 2017, the frequency and magnitude of floods in relation to climate change is also a major concern especially in the Hindu Kush region (Rahman & Khan, 2013). In the HK region, semi-annual flooding causes damages to lives, standing crops, and various infrastructures. In the history of Pakistan, a severe flood in 2010 hit the eastern Hindu Kush region and caused more than 400 fatalities and ruined the physical infrastructure including buildings, roads, bridges, and water supply schemes. The aim of this study is to determine the impact of climate change on flood factors and the extent of damages in the Hindu Kush region.

Materials and Methods

For this study, data were collected both from primary and secondary sources (Figure 3). Primary data were gathered from the flood-affected population, concerned government officials and field observations, focused group discussion (FGDs) with the community elders, and field observations. During field surveys, flood victims were asked about the possible causes of floods and the extent of associated damages. Similarly, officials of the line agencies were also interviewed to find the policy gap between the community vulnerability and their strengths.

Secondary data regarding temperature, rainfall, snowfall, and river discharge were acquired from Pakistan Meteorological Department (PMD), Peshawar. Past flood damage data were collected from Provincial Disaster Management Authority (PDMA) Peshawar. In the study region, there are seven meteorological stations: Drosh, Chitral, Dir, Timergara, Kalam, Saidu, and Malam Jabba. In order to quantify the climate change, the temperature time-series data for all the meteorological stations were analyzed by applying Mann-Kendall Trend Model (MKTM) with 95% confidence level, whereas the rainfall data were tested using auto regressive integrated moving averages (ARIMA). Similarly, the snowfall data were calculated and analyzed using the so-called least-square technique, while correlation of the climate change parameters against river discharge was analyzed using Pearson Product Movement Correlation Coefficient (PPMCC). All the meteorological stations possess different time periods of data: Chitral holds 51 years of data, Dir have 47 annotations, Drosh has 32, Saidu has 40, Kalam and Malam Jabba has 11, while Timergara has 7 years of time-series data.

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush RegionClick to view larger

Figure 1. Location of the Hindu Kush region showing administrative regions and drainage pattern.

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush RegionClick to view larger

Figure 2. Digital terrain of Hindu Kush region modified after Rahman and Khan (2013).

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush RegionClick to view larger

Figure 3. Research process.

In climate change literature, MKTM has been widely used to quantify the temperature fluctuations and its trends non-parametrically (Modarres & Silva, 2007). These models do not need normal distribution in data (Douglas et al., 2000). In order to reject null hypothesis (Ho), the probability value (p-value) must be less than the significance level (0.05), which indicates the occurring trend in temperature data. Contrary, if the probability value is equal to 0.05, then Ho will be accepted indicating no trend in data. The ARIMA model is developed by Box and Jenkins (1970) and is an applied approach; it plays a vital role in prediction and analyzing the rainfall data (Zhang et al., 2011; Tseng & Tzeng, 2002). Additionally, GIS was also applied on all the met data for preparing spatial maps.

Finally, the results were visualized in the form of maps, statistical diagrams, and tables. And then the results were illustrated and analyzed in the light socioeconomic and physical environment of Hindu Kush region.

Impact of Climate Change on Flood Factors

There are varieties of factors that generate floods in the Hindu Kush region. For detailed analysis, causative factors that trigger floods have been classified into physical and human- induced factors. In the Hindu Kush region, physical factors are contributing much to the genesis of floods. However, in certain cases, human-induced factors further intensify the recurrent flood events. In the headwater region, flash-flood characteristics dominate, whereas in the lower reaches riverine floods are dominant.

Physical Factors

In the HK region, almost every summer brings monsoonal rainfall and high temperature, which join hands with heavy melting of glaciers and snow accelerate discharge in the river system.

Temperature Fluctuations and Floods

In the HK region, in valleys there are warm/hot summers and cool/cold winters, whereas in the high altitudes, there is warm/mild summer and cool/cold winter. For the seven met stations (Chitral, Dir, Drosh, Malam Jabba, Saidu, Kalam, and Timergara), mean monthly maximum temperature were collected from the year of establishment of each met station, and the same was analyzed accordingly. It is, therefore, true that time-series temperature data varies from station to station as Chitral holds 51 years of data, Dir possesses 48 years, Drosh has 33 years, Saidu has 41 years, Malam Jabba has 11 years, Kalam has 11 years, and Timergara has 7 observations only. In the study region, Saidu is the station where the highest temperatures are recorded, while the lowest have been recorded at Malam Jabba. The ever-highest maximum temperature was recorded at Saidu (39.9oC) in 1985 followed by Drosh at 39.2oC and Timergara at 39oC. In all the met stations, the variation in mean monthly highs is mainly due to its altitude, forest cover, and rainfall pattern. The trend in temperature for a specific place changes with altitudes. Timergara, with its low elevation (460 m), experiences high temperatures, while the rest of the meteorological stations have registered comparatively low temperatures. Contrary to this, Malam Jabba, with its high elevation (2,591 m), recorded the lowest mean monthly maximum temperature (23.4oC).

At Kalam, July is the warmest month with the highest recorded mean monthly maximum temperature of 25oC, while at Malam Jabba, the highest recorded maximum temperature is 21.9oC in June. Similarly, rest of the met stations (Drosh, Timergara, and Chitral) show diverse mean monthly maximum temperatures depending upon rainfall pattern and elevation. By applying MKTM, trends in mean monthly maximum temperature are recorded at Saidu with p-value=0.032 and Dir with p-value=0.003. At both the met stations (Dir, Saidu), the p-value was recorded less than the 0.05%, which indicate that there is trend in mean maximum temperature (Table 1).

Table 1. MKTM Results of Mean Monthly Maximum Temperature for the Entire Seven Met Stations (in Degree Celsius)

Met Station

Mann-Kendall Statistics

Kendall’s Tau

Variance (S)

P-Value (Two-Tailed Test)

Model Interpretation

Timergara

–5.0

–0.3

0.0

0.4

Accept Ho (NST)

Saidu

164.0

0.2

0.0

0.03

Reject Ho (ST)

Malam Jabba

2.0

0.2

0.0

0.8

Accept Ho (NST)

Kalam

4.0

0.1

0.0

0.7

Accept Ho (NST)

Dir

237.0

0.3

6240.0

0.003

Reject Ho (MST)

Drosh

99.0

0.2

3120.3

0.07

Accept Ho (NST)

Chitral

364.0

0.3

12378.6

0.001

Reject Ho (MST)

Notes: *No Significant Trend (NST);

**Significant Trend (ST);

***More Significant Trend (MST).

In the headwater region of River Swat, River Panjkora, and River Chitral high altitudes have low temperatures that provide opportunity for snowfall, and as a result it contributes a positive regime to the glaciers. In the study region, every year the upper catchment area of all the rivers receives ample amount of snow during winter and spring. Such snow accumulation together with the melting of over 3,000 small and large glaciers, increases river discharge. In summer, melting of accumulated glaciers and snow also combines with monsoon rain that further intensifies river discharge in the Hindu Kush region: as a result, this increases the potential of recurrent floods.

The analysis demonstrates that among selected seven met stations, in January mean monthly minimum temperature is recorded at Kalam (–8oC), followed by Malam Jabba with a mean monthly minimum temperature of –4.7oC. After applying MKTM, temperature trends have been detected at Timergara and Saidu met stations, where the p-value is less than the significance level (Table 2). After putting the values in MKTD, the lowest-ever recorded temperature was recorded at Kalam met station (i.e., –18oC, followed by Malam Jabba with –13oC). The analysis reveals that temperature decreases from lower to higher altitudes and due to low temperature in winter, the snowfall and area under snow cover increases, which ultimately enhances river discharge in summer, which most often leads to flood events. At the met stations of Malam Jabba, Kalam, Drosh, and Chitral in winter, the monthly minimum temperature falls below freezing point due to the highland climate. This also increases the probability of precipitation in the form of snow, which later on in summer becomes part of the runoff.

Table 2. MKTM Results of Mean Monthly Minimum Temperature for all the Seven Met Stations (in Degree Celsius)

Met Station

Mann-Kendall Statistics

Kendall’s Tau

Variance (S)

P-Value (Two-Tailed Test)

Model Interpretation

Timergara

–10.0

–1.0

0.0

0.01

Reject Ho (MST)**

Saidu

–176.0

–0.2

0.0

0.02

Reject Ho (MST)

Malam Jabba

–4.0

–0.4

0.0

0.48

Accept Ho (NST)

Kalam

0.0

0.00

0.0

0.90

Accept Ho (NST)

Dir

–104.0

–0.1

6253.3

0.19

Accept Ho (NST)

Drosh

16.0

0.03

3097.3

0.78

Accept Ho (NST)

Chitral

–169.0

–0.1

12229.6

0.12

Accept Ho (NST)

Notes: *No Significant Trend (NSTD);

**Significant Trend (STD);

***Most Significant Trend (MSTD).

In addition to mean monthly maximum and mean monthly minimum temperature, MKTM was also applied to monthly normal temperature. The analysis reveals that the met station of Malam Jabba holds the highest monthly normal temperature in January, followed by the Kalam met station. Whereas rest of the met stations showed fluctuation in monthly temperature norms. After applying MKTM to monthly normal temperature of all the seven met stations, no prominent trend is detected (Table 3). The probability value was mostly found greater than the standard level, as a result, Ho is accepted indicating that there is no temperature variation in terms of monthly normal temperature.

Table 3. MKTM Results of Monthly Normal Temperature for all the Met Stations

Met Station

Mann-Kendall Statistics

Kendall’s Tau

Variance (S)

P-Value (Two-Tailed Test)

Model Interpretation

Chitral

152.0

0.1

12635.3

0.17

Accept Ho (NSTD)

Drosh

71.0

0.1

3134.3

0.21

Accept Ho (NSTD)

Dir

136.0

0.1

6326.0

0.09

Accept Ho (NSTD)

Kalam

0.0

0.0

0.0

0.90

Accept Ho (NSTD)

Malam Jabba

0.0

0.0

14.6

1.00

Accept Ho (NSTD)

Saidu

20.0

0.03

0.00

0.80

Accept Ho (NSTD)

Timergara

−9.0

−0.6

0.0

0.13

Accept Ho (NSTD)

Note: *No Significant Trend (NSTD).

Rainfall and Floods

In the HK region, rainfall mostly occurs in the two major seasons, winter and summer. During winter, the region gets rainfall from Western depressions, whereas in summer rainfall is from monsoon spells. Among the met stations, Drosh and Chitral get the lowest annual rainfall. In the HK region, Dir, Malam Jabba, Kalam, and Saidu receives maximum rainfall during the month of March and identified as the wettest month. Monsoonal rainfall is predominantly concentrated in summer months (July and August) in the met stations of Dir, Malam Jabba, and Saidu. The rest of the year (i.e., May, June, October, November, and December) is comparatively dry. In the study region, Malam Jabba is the humid station and receives an average annual rainfall of over 1,640 mm, followed by the met station Dir (1,362 mm), whereas the remaining stations receive less rainfall. This indicates that in the drainage basin of the River Swat, Malam Jabba, and Kalam are the wettest stations. Further, every year plenty of snowfall also occurs in these areas. Similarly, the humid station Dir lies in the headwater region of River Panjkora. Therefore, during summer months, heavy rainfall, together with the accelerated snow and glacier melting, increases discharge in all the rivers originating from the Hindu Kush region and causes floods.

The analysis further indicates that the rainfall generally varies from one meteorological station to another. After applying the least-square techniques, the trend line on annual rainfall illustrates rising trends. Likewise, at Saidu met station, the average annual rainfall is 1,051 mm. At Chitral met station (51 years of data), the mean annual rainfall is 458 mm, and it was confirmed from the analysis that an increasing (positive) trend was found over time. At Drosh for 33 annotations (1982–2014), 568 mm mean annual rainfall was received, and it was in 1992 when the highest-ever rainfall of 867 mm occurred while the lowest-ever (349 mm) was in 2002. This indicates that a high anomaly is recorded in annual rainfall at Drosh.

Based on rainfall time-series data, a linear trend line has been drawn. In case of annual rainfall, this declining trend has been conspicuous since 1994. At Kalam, for the 11 observations, there is a high variability in rainfall occurrence and the mean annual rainfall (1,039 mm), whereas the highest-ever recorded rainfall is 1,351 mm in 2005. The analysis further revealed that on average annual rainfall, a trend line is showing increasing rainfall in Kalam. Similarly, at Malam Jabba, the mean annual rainfall for 11 years is 1,648 mm, while the highest-ever rainfall (2081 mm) took place in 2006. Similarly, Timergara met station received highest-ever annual rainfall of 1,059 mm in 2010. It was found from the analysis that at Timergara a decreasing trend in annual rainfall had been recorded.

By applying the ARIMA model to Dir met station using average annual rainfall data, the rainfall has shown to fluctuate. In the case of integrated moving averages, the highest ARIMA (2,028 mm) was noticed in 1987, however the lowest at 861 mm was recorded in 1972. For Saidu met station, in terms of ARIMA, the highest ARIMA value of 1581 mm was recorded in 1992 and the lowest (531 mm) was registered in 1974. In the same way, at Chitral meteorological station, subsequently, the analysis discloses that the highest ARIMA in terms of moving averages (729 mm) was recorded in 1973 while the lowest (136 mm) was noted in 1964. In the case of Kalam met station, the lowest ARIMA was detected in 2010, and the anomaly was great in the same year. The same can be confirmed from the worst flood of 2010.

Furthermore, after applying ARIMA to the rainfall data of Malam Jabba met station, the highest ARIMA (2,032 mm) is recorded in 2007. This is the highest ARIMA among all the seven met stations. At met station Timergara, the ARIMA values also exhibit variability in terms of moving averages and anomalies. At Drosh met station, based on annual rainfall time-series data for the entire 33 observations, the analysis disclosed a great variability in terms of moving averages. The highest ARIMA was in 1993, while the lowest was identified in 2003. The ARIMA results in all of the met stations illustrate that there is wide fluctuation in mean annual rainfall.

The analysis of anomalies in time-series climatic data plays an important role in studying weather pattern, trends, and anomalies in climate change scenarios. Generally, an anomaly is the difference between the actual annual rainfall and the ARIMA-predicted moving average values. If the resultant anomaly value is more than the ARIMA value, the anomaly will be positive; and in the case of the figure being less than the ARIMA predicted value, the resultant anomaly will be negative. So according to this rule, the ARIMA PDQ model has been applied to Dir station, where the highest-ever positive anomaly (873 mm) was detected in 1986. Contrary to this, the lowest-ever negative anomaly (–21 mm) was identified in 1971 (Figure 4). The overall analysis revealed that at Dir there is a positive anomaly trend based on annual rainfall. Such increasing trends in annual rainfall and positive anomalies is an indication of accelerated intensity of discharge in the River Panjkora.

Similarly, the ARIMA model has been applied to Saidu met station. It was found from the analysis that a negative anomaly was observed in average annual rainfall. It is a low-lying met station and receives comparatively less rainfall. The data further shows that there is an increasing trend during the period from 1974 to 2014. At Chitral, using the ARIMA geo-statistical model, an increasing trend in annual rainfall anomaly has been observed. In 1972, the highest-ever positive anomaly of 565 mm was noted, whereas the lowest-ever negative anomaly (222 mm) was detected in 2001. The analysis further demonstrates that at Drosh met station, after applying the ARIMA model, the highest positive anomaly of 416 mm was registered in 1986. Whereas, the lowest-ever negative anomaly of –299 mm was noted in 1993. In the HK region, out of a total of seven meteorological stations, Malam Jabba, Kalam, and Timergara possess short time-period data. After applying the ARIMA model to these stations, a great fluctuation has been detected in terms of moving averages as well as anomalies. The overall calculations and results concerning rainfall data express an increasing trend in negative anomalies, which trigger floods in the region.

Snowfall and Floods

Generally, the meteorological stations located in the HK region receive ample precipitation. Among them, Malam Jabba, Kalam, and Dir met stations receive more precipitation in the winter season in the form of snow. The met stations that lie on the eastern side such as Saidu, Malam Jabba, and Kalam receive comparatively more rainfall during the monsoon season. However, the met stations located on the western side (Timergara, Dir) received more rainfall in winter and spring months, as it is close to the contact region. Such snow cover encourages river discharge in summer months (June, July) due to high temperature and consequently escalates discharge in rivers. It is calculated from the analysis that Kalam and Malam Jabba met stations received a comparatively greater amount of snow. At Kalam, the highest-ever snowfall was noted in 2009 (782 cm), and a super flood event happens in Swat Valley and the downstream surrounding areas. Compared to these two met stations (Malam Jabba, Kalam), Dir receives less snow; the highest-ever snowfall (102 cm) occurred in 2006. In the study region, river discharge varies from season to season. The river discharge increase in summer due to monsoonal rainfall and high temperature increases the melting of snow and glaciers. In the HK region, heavy rainfall and along with melting snow increases the potentials of flood events.

In the face of climate change, a significant correlation between rainfall patterns, trends in temperature, and resultant peaks in river discharge have been recorded. It was found that there is a rising trend in temperature, which leads to early and rapid melting of glaciers and snow in the headwater region. The analysis reveals that since the 1980s, radical changes in behavior of numerous valley glaciers have been noted. Similarly, a fluctuation in the amount of snowfall occurrences together with its timing and seasonality has been recorded. In addition, the spatial and temporal scales of violent weather events have been grown during the past 30 years.

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush RegionClick to view larger

Figure 4. ARIMA detected anomaly in annual rainfall, (a) Dir, (b) Saidu, (c) Chitral, (d) Drosh, (e) Kalam, (f) Malam Jabba, (g) Timergara.

Human-Induced Factors

Along with the physical factors there are some flood intensifying factors that accelerate floods and incur damages to the area. In the Hindu Kush region, human encroachments over the active flood plain, population growth, deforestation, and overgrazing in the catchment area are some of the key anthropogenic factors. Such changes in water regimes, including the frequent but substantial increase in heavy precipitation events, rapid and early melting of snow/glaciers in the headwater region, siltation in active channels, excessive deforestation in the past three decades, and human encroachments onto the active flood channel have further escalated the flooding events.

Forest cover plays a key role in mitigating flood disasters, while deforestation accelerates floods (Haeusler, Schnurr, & Fischer, 2000). According to FAO, in the Hindu Kush region trend of annual forest cover change during 1990 to 2000 was –1.8%, while during 2000–2008 it was –2.2 percent (Ali, 2008; Figure 5). Analysis revealed that since 1991 a drastic change has occurred in land use and land cover in the Hindu Kush region. The forest cover has been diminished by about 4%, and the built-up area has increased from 0.42% of the total area to 1.16% in 2015.

Field survey together with the FGDs reveals that the main factors behind the deforestation are population pressure, supply of timber to the market, clearing of forest for agriculture, and built-up area. Similarly, the growing population also demands institutions, resource base, and infrastructure. It is reported that during the recent militancy in the province of Khyber Pakhtunkhwa, extensive forest cutting has taken place in the thickly forested districts of Swat, Shangla, Dir upper, and Dir lower (Ali, 2008; Rahman & Khan, 2013). While in certain cases the timber mafia have also been blamed for ruthless cutting of forest. It is therefore, in the study area, the rate of deforestation increased significantly mainly at the cost of revenue generation and extension of amenities to the dwellers.

Field survey, FGDs, and interviews with the officials of the forest department revealed that in the Hindu Kush region deforestation and overgrazing are the major flood intensifying factors. Thus, this rapid deforestation and overgrazing have seriously affected the fluvial processes of almost all the river systems originating from the Hindu Kush region (Khan & Rahman, 2006; Ali, 2008; Rahman, 2010).There is also a strong correlation between deforestation and flood events. Consequently, this phenomenon has reduced the water absorption capacity and contributed much to intensifying the frequency and magnitude of floods in the region.

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush RegionClick to view larger

Figure 5. Land use/land cover change in the drainage basin of river Swat, Hindu Kush Region.

In the study area, population is increasing at a rapid pace (2.6% per annum), which put tremendous pressure on the land resources (GoP, 2000a, 2000b, 2000c, 2000d). Because of this, both the forest and farmland are converted into built-up areas. The area under concrete structures, consisting of, for example, houses, paved streets, roads, commercial areas, and other buildings, have been increased but at the cost of degrading either agricultural or forest land. As a result of all these developments, the absorption capacity of rainwater has decreased drastically, and most of which quickly becomes part of surface runoff. Thus, increase in the built-up area has a direct relationship with increasing river runoff and intensifying floods in the Hindu Kush region.

With the rapid pace of population growth, the demand for shelter and other infrastructures also increased. Poor people have no choice but to encroach on the active floodplains either for agricultural or infrastructural purposes. The river course narrows down and in effect reduces the carrying capacity. Therefore, during intense rainfall, the water overflows the levee and ultimately causes floods. During field survey, it was observed that human encroachment has largely contributed to the occurrence of 2010 floods. Therefore, there is a need of policy response for the land-use regulation to check encroachments onto the channel and watershed management as a flood abatement measure to reduce the risk of future mega flood events.

In the Hindu Kush region, there are hundreds of bridges over the streams and rivers. Torrential rainfall, cloud bursting, lateral erosion, and heavy runoff often uproot the trees especially in the upstream areas (Rahman & Khan, 2013). Such tall trees when picked up by gushing water pass through the narrow bottleneck of bridges. The trees usually blocked the channel. As a result, a temporary dam develops behind the bridges, which constantly puts pressure on the bridge. Consequently, the bursting of temporary dams causes the bridge to be compromised, and thus the stored water would inundate the downstream areas. Bridges are constructed to connect physically separated areas. In the 2010 flood, several bridges were damaged. After destruction of Chakdara Bridge, a major supply line of the three districts (in the Hindu Kush region) was for a month totally cut off from rest of the country. This in turn led to a rise in the cost of food items in the detached areas. Similarly, due to the phenomenon of bursting temporary dams, peak discharges over the hydro-gauging stations have been recorded. During the 2010 flood, this was observed as a major triggering factor of floods in the entire Hindu Kush region.

In the Hindu Kush region, several gauging stations are on the rivers Swat, Panjkora, and Chitral. Data was obtained from the Irrigation and Drainage Authority Peshawar and Surface Water Hydrology, WAPDA Lahore. Discharge data of Swat at Amandara headworks (located upstream in the Panjkora confluence) and Munda headworks (a point downstream from the confluence of the River Swat and Panjkora) were obtained since their establishment. The River Swat and Panjkora have an irregular peak flow with high interannual variability in runoff. The mean monthly discharge at Amandara continues to rise from March till August, and then it starts gradually falling. The discharge data reveals that peak runoff at Amandara has been recorded during July and August and also shows unusual high discharge during 1983, 1992, 1995, 2001, and 2010 (Figure 6). Nevertheless, on August 9, 1992 (3,588m3/s) and on July 25, 1995 (3,630m3/s) maximum discharge was recorded at Amandara headworks. However, the 2010 flood has broken all the previous records and even destroyed Amandara headworks. In July 2010, the discharge at Amandara exceeded 5,663m3/s (Rahman & Khan, 2013).

Impact of Climate Change on Flood Factors and Extent of Damages in the Hindu Kush RegionClick to view larger

Figure 6. (a) Swat River, highest discharge recorded at Amandara from 1983 to 2011 (b) Swat River, maximum discharge recorded at Munda headworks from 1929 to 2011.

The discharge recorded at Munda shows that the unusually high runoff was recorded in 1929, 1962, 1966, and 2010 (Figure 6). The analysis indicates that in the past 80 years, the highest recorded discharge at Munda was 4,502m3/s in 1929: this is the highest potential capacity of Munda headworks. At Munda headworks more than 8,495m3/s was recorded on July 29, 2010, which was the ever-highest discharge during the past 80 years. This heavy deluge of water also destroyed the Munda headworks. In 2010, the discharge at all the gauging stations was recorded as the highest ever and caused the century’s worst flood. The analysis further reveals that during the study period there was an increasing trend in river discharge recorded both at Amandara and Munda headworks.

Therefore, there is the need for policy response for the land-use regulation to check encroachments onto the channel and watershed management as a flood abatement measure to reduce the risk of future mega flood events.

Extent of Flood Damages

Analysis revealed that the 2010 flood affected the entire Hindu Kush Region while floods in the year 2015 and 2016 affected the Panjkora basin and the upper reaches of the River Kabul basin. It has undercut the foundation of the buildings, roads, and irrigation channels; it has also eroded topsoil and changed river morphology. Similarly, the flood resulted in human and animal life losses. Within four selected sample districts—namely Upper Dir, Lower Dir, Swat, and Chitral—the extent of damages was determined through extensive field surveys, observations, and interviews. The extent of damages in various sectors is given in the following section.

Floods have caused the loss of many human and animal lives. Most of the human casualties were in the upper catchment of River Kabul, River Swat, and River Panjkora areas where the gradient is more than 30 degrees. In Swat and Upper Dir, the impact of flooding on human life was highest, with total deaths at 95 and 94, respectively. Most of the victims were working in their fields near the river. In the year 2015, flash floods resulted in 17 injuries and 50 fatalities, with highest number in the district of Chital. These deaths and injuries occurred because of the lack of any warning or emergency systems. In the eastern Hindu Kush, total fatalities and injuries have been 781 and 342, respectively, since 2010 (Table 4).

Table 4. Injuries and Fatalities Caused by Flood in Eastern Hindu Kush

Flood District

2010

2015

2016

Injuries

Fatalities

Injuries

Fatalities

Injuries

Fatalities

Dir Upper

424

94

5

13

14

12

Dir Lower

30

6

4

7

Swat

207

95

1

1

17

17

Chitral

62

21

11

36

6

40

Total

723

216

17

50

41

76

Most of the human settlements are located near the river on the left and right banks. The majority of the people are living in mud-and-stone houses in the upper catchment of all rivers in the study region, while the lower zone of the catchment is characterized by multistory concrete residential and non-residential buildings. Most of the flood-damaged buildings are located close to the river because they were highly vulnerable. The load and speed of the flood have played an important role in the destruction of buildings. In the 2010 flood disaster, there were more badly damaged houses in in Swat (14,460) followed by Upper Dir (792), Chitral (365), and Lower Dir (260). In 2015, Chitral was severely affected by flash floods with a number of partially damaged (683) and totally destroyed (803) houses in the region (Table 5). Similarly, the extent of damages to non-residential buildings was also severe. The largest number of damaged buildings was in Swat followed by Upper Dir, Lower Dir, and Chitral. The number of commercial buildings that were damaged is high because of their location near the rivers. So far floods have also damaged non-residential building including commercial (312), educational (140), and religious (100) since 2010 (Table 6).

Table 5. Damages to Houses in Eastern Hindu Kush (2010–2016)

Flood District

2010

2015

2016

PD

CD*

PD

CD

PD

CD

Dir Upper

980

792

2

4

0

104

Dir Lower

603

260

2

2

Swat

14460

26

17

0

172

Chitral

165

365

683

803

19

67

Total

1748

15877

711

826

19

345

Notes: * Partially Damaged (PD);

** Completely Damaged (CD).

Table 6. Damages to Non-Residential Buildings in Eastern Hindu Kush (2010–2016)

Buildings District

Flood 2010

Flood 2015

Commercial

Educational

Health & Govt.

Religious

Commercial

Educational

Religious

Dir Upper

40

40

19

19

Dir Lower

72

1

13

1

Swat

161

69

13

69

Chitral

17

9

22

30

2

Total

290

110

45

98

22

30

2

High drainage density is a natural obstruction to people’s movement from one village to another. Bridges and culverts have been used to fill this gap. Most of these structures were completely or partially damaged by floods. The 2010 flood damaged the highest number of bridges in Chitral (39), followed by Upper Dir (26), Swat (24), and Lower Dir (17). In the year 2015, Chitral was severely affected, with the highest number of damaged bridges (31) and culverts (27; Table 7). Peak discharge with maximum hydraulic pressure at bridging location and destructive load (including tree logs and timber) are the causative factors behind bridges being destroyed.

Roads are the only means of getting around in eastern Hindu Kush. Main roads have been constructed near rivers in the entire study region. Seasonal and perennial streams have bisected the roads at many locations because of high drainage density. Roads are strengthened by retaining walls. The floods undercut the foundation of the retaining wall where the road is near the river level: this has resulted in much damage. The total length of damaged road was 135 km (2010), 19.5 km (2015), and 4.5 km (2016; Table 8)

Water supply schemes were also severely affected by floods. Head sources of water supply schemes are located in the upstream of hill torrents. Destructive flow in the 2010 flood completely washed away the head sources of these schemes. Most of the damages were concentrated in Upper Dir (92) and Lower Dir (70). During the 2015 flood, the number of damaged water supply schemes was 205 in Chitral and 4 in Dir Upper (Table 9).

Table 7. Damages to Culverts and Bridges in Eastern Hindu Kush

Flood District

2010

2015

Bridges

Culverts

Bridges

Culverts

Dir Upper

26

22

1

12

Dir Lower

17

14

1

Swat

24

Chitral

39

31

27

Total

106

36

33

39

Table 8. Damages to Roads (KM) in Eastern Hindu Kush

Flood District

2010

2015

2016

Dir Upper

57

2

4

Dir Lower

63

0.5

0.5

Swat

2

Chitral

13

17

Total

135

19.5

4.5

Table 9. Damages to Water Supply Schemes in Eastern Hindu Kush

Flood District

2010

2015

2016

Dir Upper

92

4

7

Dir Lower

70

Swat

Chitral

205

Total

162

209

7

In the eastern Hindu Kush region, people earn their livelihood mostly from agricultural activities such as terraced farming. Cultivation fields near rivers, streams, and torrents were damaged by riverine and flash floods. In the upper zone, agricultural land near the river was eroded by flash floods, while in the lower zone land and crops were damaged by the temporary deposits in the form of sand and pebbles and erosion as well. The layer of sand and pebbles over the agricultural land ranged from 0.8 m to 2 m. Maximum cultivated land was damaged by the 2010 flood in Upper Dir (40,995 ha) followed by Swat (13,950 ha) and Lower Dir (3,436 ha). In the 2015 flood Chitral was most affected with 1,023 ha of cultivated land (Table 10). In 2010 and 2015 floods took the lives of thousands of livestock. Upper Dir had the highest losses of buffalos (80), cows (3,765), sheep (2,010), and goats (6,258). The number of chicken (15,510) lost was also the highest (Table 11).

Table 10. Damages to Agriculture Land (HA) in Eastern Hindu Kush

Flood District

2010

2015

2016

Dir Upper

40995

27

44

Dir Lower

3436

Swat

13950

Chitral

1023

Total

58381

1060

44

Table 11. Damages to Livestock in Eastern Hindu Kush

Flood District

2010

2015

Buffalos

Cow

Sheep

Goat

Chicken

Buffalos

Cow

Sheep

Goat

Chicken

Dir Upper

80

3765

2010

6258

15510

4

Dir Lower

2

6

40

1317

Swat

Chitral

43

1100

615

2000

1358

Total

82

3771

2010

6298

16827

43

1104

615

2000

1358

Conclusion

The Hindu Kush region is exposed to numerous disasters, but flooding is a recurring phenomenon. It is a humid to semi-arid climatic region and a source of hundreds of glaciers. Several rivers therefore originate from the Hindu Kush region. The analysis revealed that in the upper reaches flash flood characteristics dominate, while in the lower reaches riverine flooding is common. Both physical and human factors are generating floods. However, the contribution of climatic factor is more important than the anthropogenic factors.

This study was conducted to explore the trend in temperature and its impact on floods in the Hindu Kush (HK) region. This study focuses on temperature, rainfall, snowfall, and river discharge using the MKTM, ARIMA model, PPMC, and the least-square method. After applying MKTM to the temperature data, the analysis revealed that there are significant statistical trends occurring in the meteorological stations of Chitral, Drosh, Dir, and Saidu, which have long-term data; while no prominent temperature trend was detected in the met stations of Malam Jabba, Kalam, and Timergara. Similarly, in case of monthly minimum temperature, a significant trend occurred in the meteorological stations of Saidu and Timergara, while no trend was found for monthly normal temperature. In case of rain, it is concluded from the analysis that Malam Jabba, with an average annual rainfall of 1,647 mm, is the wettest station followed by Dir met station, (1,362 mm) whereas the rest of the met stations indicated fluctuation.

In the context of ARIMA, Saidu, Chitral, Dir, and Kalam met stations have shown increasing trends. In case of rainfall anomaly, high peaks of negative and positive anomalies were found. It is also an indication of uncertainty in rainfall. In the HK region, Malam Jabba, Kalam, and Dir receives ample precipitation in the winter season mostly in the form of snow. After melting, such snow encourages runoff in June, July, and August. Therefore, due to this unexpected runoff pattern, there are more chances of flood events in the future.

It is concluded from the analysis that heavy and prolonged rainfall in summer is a major factor responsible for the genesis of flood disaster. In addition to this, excessive melting of snow and glacier, aggradation of river bed by consistent sedimentations, and cloud bursting are the major physical factors responsible for generation of floods. There are some human factors as well that further intensify the probability of floods in the Hindu Kush region. Out of these, human encroachments onto the active flood plain, bursting of temporary dams behind the bridges, clearing of forest and agricultural land for infrastructure, expansion of the area under impermeable soil, deforestation, and overgrazing in the catchment area are some of the key anthropogenic factors. These factors have also exacerbated the extent of damages. Therefore, there is need of policy reform so the land-use regulation can better check encroachments onto the channel and watershed management as a flood abatement measure to reduce the risk of future mega-flood events. The analysis further indicates that there is a constant fluctuation in both the rainfall and river runoff, which is a clear indication of climate change.

References

Ali, A. M. S. (2007). September 2004 flood event in southwestern Bangladesh: A study of its nature, causes, and human perception and adjustments to a new hazard. Natural Hazards, 40, 89–111.Find this resource:

Ali, H. B. (2008, October 1). Peshawar: Swat forests at the mercy of militants. Dawn News.Find this resource:

Benestad, R. E., & Haugen, J. E. (2007). On complex extremes: Flood hazards and combined high spring-time precipitation and temperature in Norway. Climate Change 85(3), 381–406.Find this resource:

Briffa, K. R., Jones, P. D., Schweingruber, F. H., Shiyatov, S. G., & Cook, E. R. (1995). Unusual twentieth-century summer warmth in a 1,000-year temperature record from Siberia. Nature, 376(6536), 156–159.Find this resource:

Chang H., Franczyk, J., & Kim, C. (2009) What is responsible for increasing flood risks? The case of Gangwon Province, Korea. Natural Hazards, 48(3), 339–354.Find this resource:

Douglas, E. M., Vogel, R. M., & Kroll, C. N. (2000). Trends in floods and low flows in the United States: Impact of spatial correlation. Journal of Hydrology, 240, 90–105.Find this resource:

Douglas, L. (2011). Global weather highlights 2010: Flooding, heatwaves, and fires. Weatherwise, 64(3), 21–28.Find this resource:

Government of NWFP. (2006). Environmental profile of NWFP. Environmental Protection Agency, Government of NWFP, Peshawar-Pakistan.Find this resource:

Government of Pakistan (GoP). (2000a). District census report of Swat, 1998. Population census organization of Pakistan, Islamabad.Find this resource:

Government of Pakistan (GoP). (2000b). District census reports of Upper Dir, 1998. Population census organization of Pakistan, Islamabad.Find this resource:

Government of Pakistan (GoP). (2000c). District census reports of Lower Dir, 1998. Population census organization of Pakistan, Islamabad.Find this resource:

Government of Pakistan (GoP). (2000d). District census reports of Chitral, 1998. Population census organization of Pakistan, Islamabad.Find this resource:

Gupta, V., & Sah, M. P. (2008). Impact of the trans-Himalayan landslide lake outburst flood (LLOF) in the Satluj catchment, Himachal Pradesh, India. Natural Hazards, 45, 379–390.Find this resource:

Haeusler, T., Schnurr, J., & Fischer, K. (2000). Provincial forest resource inventory (PFRI) North West Frontier Province. Pakistan Forest Institute, Peshawar-Pakistan.Find this resource:

Hunter, N. M., Horritt, M. S., Bates, P. D., Wilson, M. D., & Werner, M. G. F. (2005). An adaptive time step solution for raster-based storage cell modelling of floodplain inundation. Advances in Water Resources, 28, 975–991.Find this resource:

IPCC. (2014). Core Writing Team. In R. K. Pachauri & L. A. Meyer (Eds.), Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (p. 151). Geneva, Switzerland: IPCC.Find this resource:

Khan, A. N., Rahman, A., & Ali, B. (2008). Adjustment to riverbank erosion hazard: A case study of Village Chalyar, Swat Valley, Pakistan. Pakistan Journal of Geography, 18 (1–2), 46–59.Find this resource:

Khan, A. N., & Rahman, A. (2006). Landslide hazards in the mountainous region of Pakistan. Pakistan Journal Geography, 16(1–2), 38–51.Find this resource:

Khan, F. K. (2003). Geography of Pakistan: Population, economy and environment. Karachi, Pakistan: Oxford University Press.Find this resource:

Lehner, B., Ll, P. D., Alcamo, J., Henrichs, T., & Frank, K. (2006). Estimating the impact of global change on flood and drought risks in Europe: A continental, integrated analysis. Climate Change, 75, 273–299.Find this resource:

Mahmood, S., Khan, A. H., & Mayo, S. M. (2016a). Exploring underlying causes and assessing damages of 2010 flash flood in the upper zone of Panjkora River. Natural Hazards, 83(2), 1213–1227.Find this resource:

Mahmood, S., Khan, A. H., & Ullah, S. (2016b). Assessment of 2010 flash flood causes and associated damages in Dir Valley, Khyber Pakhtunkhwa Pakistan. International Journal of Disaster Risk Reduction, 16, 215–223.Find this resource:

Mahmood, S. A., Shahzad, F., & Gloaguen, R. (2009). Remote sensing analysis of quaternary deformation using river networks in Hindukush region. In Geoscience and Remote Sensing Symposium, 2009 IEEE International, IGARSS August 2009 (Vol. 2). Freiberg, Germany: Remote Sensing Group, Tech. University Bergakad.Find this resource:

Mazzorana, B., Hübl, J., & Fuchs, S. (2009). Improving risk assessment by defining consistent and reliable system scenarios. Natural Hazards and Earth System Sciences, 9(1), 145–159.Find this resource:

Mirza, M. M. Q. (2003). Three recent extreme floods in Bangladesh: A hydro-meteorological analysis. Natural Hazards, 28(1), 35–64.Find this resource:

Modarres, R., & Silva, V. P. R. (2007). Rainfall trends in arid and semi-arid regions of Iran. Journal of Arid Environment, 70, 344–355.Find this resource:

Montz, B. E., Gruntfest, E. (2002). Flash flood mitigation: Recommendations for research and applications. Environmental Hazards, 4, 15–22.Find this resource:

Nathan, F. (2008). Risk perception, risk management and vulnerability to landslides in the hill slopes in the city of La Paz, Bolivia. A preliminary statement. Disasters, 32(3), 337–357.Find this resource:

Nicholson, S. E. (1994). Recent rainfall fluctuations in Africa and their relationship to past conditions over the continent. The Holocene, 4(2), 121–131.Find this resource:

Onoz, B., & Bayazit, M. (2003). The power of statistical tests for trend detection. Turkish Journal of Engineering and Environmental Sciences, 27(4), 247–251.Find this resource:

Program Adaptation to Climate Change (PACC). (2010). Heat wave in Asia beating his record of 53.50C temperature. Program of Adaptation to Climate Change—PACCPERU, Lima: Peru.Find this resource:

Qamer, F. M., Abbas, S., Saleem, R., Shehzad, K., Ali, H., & Gilani, H. (2012) Forest cover change assessment in conflict-affected areas of northwest Pakistan: The case of Swat and Shangla districts. Journal of Mountain Science, 9, 297–306.Find this resource:

Rahman, A., & Khan, A. N. (2011). Analysis of 2010-flood causes and associated socio-economic damages in Hindukush Region. Natural Hazards, 59, 1239–1260.Find this resource:

Rahman, A., Khan, A. N., Collins, A. E., & Qazi, F. (2011). Causes and extent of environmental impacts of landslide hazard in the Himalayan region: A case study of Murree, Pakistan. Natural Hazards, 57(2), 413–434.Find this resource:

Rahman, A., & Khan, A. N. (2013). Analysis of 2010-flood causes, nature and magnitude in the Khyber Pakhtunkhwa, Pakistan. Natural Hazards, 66(2), 887–904.Find this resource:

Rahman, A. (2003). Effectiveness of flood hazard reduction policies: A case study of Kabul-Swat floodplain, Peshawar vale [MPhil dissertation]. University of Peshawar, Pakistan.Find this resource:

Rahman, A. (2010). Disaster risk management: Flood perspective. Saarbrücken: Germany: VDM Verlag.Find this resource:

Rahman, A., & Dawood, M. (2017). Spatio-statistical analysis of temperature fluctuation using Mann–Kendall and Sen’s slope approach. Climate Dynamics, 48(3–4), 783–797.Find this resource:

Ramos, C., & Reis E. (2002). The analysis reveals that the flow regimes of Hindu Kush rivers largely depends on the spatio-temporal variation in precipitation. Mitigation and Adaptation Strategies for Global Change, 7, 267–284.Find this resource:

Sarikaya, M. A., Bishop, M. P., Shroder, J. F., & Olsenholler, J. A. (2012). Space-based observations of eastern Hindu Kush glaciers between 1976 and 2007, Afghanistan and Pakistan. Remote Sensing Letters, 3(1), 77–84.Find this resource:

Sarikaya, M. A., Bishop, M. P., Shroder, J. F., & Ali, G. (2013). Remote-sensing assessment of glacier fluctuations in the Hindu Raj, Pakistan. International Journal of Remote Sensing, 34(11), 3968–3985.Find this resource:

Shams, F. A. (2006). Land of Pakistan. Lahore, Pakistan: Kitabistan publisher and booksellers.Find this resource:

Zhang, S., Hua, D., Meng, X., & Zhang, Y. (2011). Climate change and its driving effect on the runoff in the ‘three-river headwaters’ region. Journal of Geographical Sciences, 21(6), 963–978.Find this resource:

Shrestha, A. B., & Shrestha, S. D. (2008). Flash flood risk in the Hindu Kush-Himalayas: Causes and Management Options. In International Conference on Disasters and Development: Bridging the Gap between Theory and Practice (p. 11).Find this resource:

Syvitski, J. P., & Brakenridge, G. R. (2013). Causation and avoidance of catastrophic flooding along the Indus River, Pakistan. GSA Today, 23(1), 4–10.Find this resource:

Tariq, M. A. U. R., & Giesen, V. D. N. (2012). Floods and flood management in Pakistan. Physics and Chemistry of the Earth, 47–48, 11–20.Find this resource:

Tseng, F. M., & Tzeng, G. H. (2002). A fuzzy seasonal ARIMA model for forecasting. Fuzzy Sets and Systems, 126(3), 367–376.Find this resource:

United Nations. (2011). State of the world’s forests 2011. Rome, Italy: United Nations, Food and Agriculture Organization.Find this resource:

Zhang, Q., Liu, C. L., Xu, C. Y., & Jiang, T. (2006). Observed trends of annual maximum water level and stream flow during past 130 years in the Yangtze River basin, China. Journal of Hydrology, 324, 255–265.Find this resource:

Zhang, Q., Xu, C. Y., Zhang, Z. X., Chen, Y. D., Liu, C. L., & Lin, H. (2008). Spatial and temporal variability of precipitation maxima during 1960–2005 in the Yangtze River basin and possible association with large-scale circulation. Journal of Hydrology, 353, 215–227.Find this resource:

Zhang, S., Hua, D., Meng, X., & Zhang, Y. (2011). Climate change and its driving effect on the runoff in the ‘Three-River Headwaters’ region. Journal of Geographical Sciences, 21(6), 963–978.Find this resource: