the climate change impact on water resources of...

THE CLIMATE CHANGE IMPACT ON WATER RESOURCES OF UPPER INDUS BASIN-PAKISTAN

By

Muhammad Akhtar M.Sc (Applied Environmental Science)

Under the Supervision of

Prof. Dr. Nasir Ahmad M.Sc. (Pb), Ph.D. (U.K)

A thesis submitted in the fulfillment of requirements for the degree of Doctor of Philosophy

INSTITUTE OF GEOLOGY UNIVERSITY OF THE PUNJAB, LAHORE-PAKISTAN

2008

Dedicated to my parents

CERTIFICATE

It is hereby certified that this thesis is based on the results of modelling work carried out

by Muhammad Akhtar under my supervision. I have personally gone through all the

data/results/materials reported in the manuscript and certify their correctness/

authenticity. I further certify that the materials included in this thesis have not been used

in part or full in a manuscript already submitted or in the process of submission in

partial/complete fulfillment for the award of any other degree from any other institution.

Mr. Akhtar has fulfilled all conditions established by the University for the submission of

this dissertation and I endorse its evaluation for the award of PhD degree through the

official procedure of the University.

SUPERVISOR

0. cL- ,Nasir Ahmad, PhDProfessor

Institute of GeologyUniversity of the PunjabLahore, Pakistan

i

ABSTRACT

PRECIS (Providing REgional Climate for Impact Studies) model developed by the Hadley

Centre is applied to simulate high resolution climate change scenarios. For the present

climate, PRECIS is driven by the outputs of reanalyses ERA-40 data and HadAM3P

global climate model (GCM). For the simulation of future climate (SRES B2), the

PRECIS is nested with HadAM3P-B2 global forcing. In the present day simulations,

climatic means and interannual variability are examined and biases are identified

focusing on the most important parameters (precipitation and temperature) for

hydrological modelling. In this study, both the meteorological station observations and

results of the PRECIS RCM are used as input in the HBV hydrological model in order to

investigate the effect of PRECIS simulated precipitation and temperature on the HBV

predicted discharge in three river basins of UIB region. For this, three HBV model

experiments are designed: HBV-Met, HBV-ERA and HBV-PRECIS where HBV is driven

by meteorological station data and by the outputs from PRECIS nested with ERA-40 and

HadAM3P data respectively. The robustness and uncertainties ranges of these models are

tested. The future water resources are quantified using the two approaches of

transferring the climate change signals i.e. delta change approach and direct use of

PRECIS data. The future discharge is simulated for three stages of glacier coverage: 100

% glaciers, 50 % glaciers and 0 % glaciers.

The PRECIS is able to reproduce the spatial patterns of the observed CRU mean

temperature and precipitation. However, there are notable quantitative biases over some

regions especially over the Hindukush-Karakorum-Himalaya (HKH) region, mainly due

to the similar biases in the driving forcing. PRECIS simulations under future SRES B2

scenario indicate an increase in precipitation and temperature towards the end of 21st

century.

The calibration and validation results of the HBV model experiments show that the

performance of HBV-Met is better than the HBV-ERA and HBV-PRECIS. However, using

input data series from sources different from the data used in the model calibration shows

that HBV-ERA and HBV-PRECIS are more robust compared to HBV-Met. The Gilgit and

Astore river basins, for which discharges are depending on the preceding winter

precipitation, have higher uncertainties compared to the Hunza river basin for which the

discharge is driven by the energy inputs. The smaller uncertainties in the Hunza river

ii

basin as compared to Gilgit and Astore river basins may be because of the stable

behavior of the input temperature series compared to the precipitation series. The

robustness and uncertainty ranges of the HBV models suggest that regional climate

models may be used as input in hydrological models for climate scenarios studies.

In a changed climate, the discharge will generally increase in both HBV-PRECIS and

HBV-Met in the 100 % glacier coverage stage up to 65% and 44%, respectively. At the 50

% glacier coverage stage, the discharge is expected to reduce up to 24% as predicted by

HBV-PRECIS and up to 30% as predicted by HBV-Met model. For the 0 % glacier

coverage under climate change, a drastic decrease in water resources is forecasted by

HBV-Met is up to 96 % and by HBV-PRECIS is up to 93%. At 100 % glacier coverage,

the magnitude of flood peaks is likely to increase in the future which is an indication of

higher risk of flood problems under climate change. There are huge outliers in annual

maximum discharge simulated with HBV-Met. This shows that the prediction of

hydrological conditions through the delta change approach is not ideal in the UIB region.

HBV-PRECIS provides results on hydrological changes that are more consistent with

climate change. This shows that the climate change signals in HBV-PRECIS are

transmitted more realistically than in HBV-Met. Therefore, the direct use of RCM outputs

in a hydrological model may be an alternative in areas where the quality of observed

data is poor. The modeled changes in future discharge and changes in peak flows under

climate change are not conclusive because more research is needed to evaluate the

uncertainties in this approach. Moreover, this technique needs to be tested with other

RCMs and hydrological models preferably to river basins in other parts of the world as

well.

iii

ACKNOWLEDGEMENTS

I would like to extend my sincere thanks to my research supervisor Prof. Dr. Nasir

Ahmad (Director Institute of Geology, University of the Punjab) for his keen interest,

proficient guidance, valuable suggestions, and encouraging attitude during the course of

this research work.

I would like to thank the PRECIS team at Hadley Centre, Meteorological Office, U.K, on

providing training in PRECIS regional climate modelling system and extending

continuous help in solving day to day operational simulation problems. Special thanks are

due to David Hein who provided boundary data of different GCMs on behalf of Hadley

Centre. The river discharge data and meteorological data have been taken from Water and

Power Development Authority (WAPDA) and Pakistan Meteorological Department

(PMD), respectively. I am grateful to the scientists at Swedish Meteorological and

Hydrological Institute (SMHI) for their useful comments and valuable suggestions during

the study.

I am indebted to the PMD on granting study leave for doctoral study at the University.

Financial support extended by the Higher Education Commission under the indigenous

PhD Scholarship Scheme is most gratefully acknowledged. I am thankful to ICTP,

Trieste, Italy, for providing two months fellowship, which enhanced my modelling

capabilities. The valuable suggestions and technical skill provided by Dr. Jermy Paul

during my stay at ICTP helped improve my understanding towards modelling technique.

Suggestions and critical comments by Dr. Martijn Booij of the Twente University,

Netherlands and Dr. David Hein and Dr. Wilfran-Moufouma of Hadley Centre, U.K.

greatly improved quality of my research work.

Finally, I would like to express my heartiest gratitude to my parents, wife, sisters,

brothers and friends whose cooperation, prayers and well wishes strengthened my

confidence to endure the hardships faced during this study.

iv

LIST OF TABLES

Table 1.1 Dimensions of some large glaciers in the UIB region 6

Table 3.1 Description of PRECIS RCM experiment 28

Table 3.2 Biases in mean temperature (˚C) as simulated with the PRECIS-Had, PRECIS-ERA and HadAM3P relative to CRU reference data for different seasons and seven sub regions of figure 3.3 (Summer= April-September, Winter = October-March)

39

Table 3.3 Biases in mean precipitation (%) as simulated with the PRECIS-Had, PRECIS-ERA and HadAM3P relative to CRU reference data for different seasons and seven sub regions of figure 3.3 (Summer= April-September, Winter = October-March)

47

Table 3.4 Seasonal changes of mean temperature and precipitation under SRES B2 scenario from PRECIS in 2071-2100 over the seven sub regions relative to 1961-1990 (Summer = April-September; Winter=October-March)

54

Table 4.1 Characteristics of study area 58

Table 4.2 Temperature and precipitation during two monsoon events at selected stations

60

Table 4.3 Biases in mean temperature (˚C) as simulated with PRECIS RCMs relative to CRU reference data for different seasons and river basins (Winter =October-March; Summer =April-September)

63

Table 4.4 Biases in precipitation (%) as simulated with PRECIS RCMs relative to CRU reference data for different seasons and river basins (Winter = October- March; Summer = April-September)

63

Table 4.5 Values and range of important parameters found in different studies using HBV model

75

Table 4.6 Parameter values for HBV for three river basins with three different input data sets

75

Table 4.7 Performance of three HBV models during calibration and validation periods in different river basins

76

Table 4.8 Efficiency Y of three HBV models using data sources different from the calibration sources during the hydrological years 1985 and 1986 in different river basins. The values of absolute relative deviations (ARD) are given in parentheses. The italic values indicate efficiency Y during calibration

86

Table 5.1 Seasonal changes of mean temperature and precipitation under PRECIS simulated SRES B2 scenario for the period 2071-2100 over three river basins relative to the period 1961-1990 (Summer = April-September; Winter=October-March)

94

v

Table 5.2 Mean relative change in future discharge (2071-2100) in a changed SRES B2 climate relative to the present discharge (1961-1990) for three glaciations stages and for three river basins

102

Table 5.3 Characteristics of future annual maximum discharge simulated by two HBV models in a changed SRES B2 climate for the three glaciations stages and for three river basins. The values in parentheses are future annual maximum discharge with outliers

106

vi

LIST OF FIGURES

Figure 1.1 Indus river basin 4

Figure 1.2 Mean annual hydrograph at different locations at the Indus river including some headwater tributaries

5

Figure 3.1 Topography of selected domain (a) Topography of the global climate model (HadAM3P), (b) Topography of the regional climate model (PRECIS), (c) Topography of the GTOPO30 2MIN DEM and (d) Deviation of PRECIS RCM topography from GCM topography

25

Figure 3.2 PRECIS RCM domain for experiments at 50 x 50 km resolution 27

Figure 3.3 Sub regions used for more detailed analysis of the PRECIS RCM fields. Region 1 (Afghanistan), Region 2 (Southern Pakistan and Rajasthan), Region 3 (Hindu Kush-Karakorum- western Himalaya), Region 4 (Central Pakistan and Northwestern India), Region 5 (Tibetan Plateau), Region 6 (Central Himalaya) and Region 7 (Central India)

29

Figure 3.4 Mean seal level pressure (MSLP) for the period 1981-1990 for May-June (MJ) season in (a) NCEP reanalyses data, (b) ERA-40 Reanalyses data, (c) HadAM3P GCM and (d) PRECIS-Had and (e) PRECIS-ERA

31

Figure 3.5 Observed and simulated (baseline) patterns of annual temperature (˚C) for (a) CRU data, (b) HadAM3P, (c) PRECIS-Had and (d) PRECIS-ERA

34

Figure 3.6 Bias of annual temperature (˚C) for (a) PRECIS-Had and (b) PRECIS-ERA with respect to CRU data

36

Figure 3.7 Observed and simulated (PRECIS-Had and HadAM3P) annual cycle of temperature averaged over the seven sub regions of figure 3.3

37

Figure 3.8 Observed and simulated (PRECIS-ERA) annual cycle of temperature averaged over the seven sub regions of figure 3.3

38

Figure 3.9 Observed (CRU) and simulated (PRECIS-Had and HadAM3P) seasonal temperature standard deviation averaged over the seven subregions of figure 3.3

39

Figure 3.10 Observed and simulated (baseline) patterns of annual precipitation (mm/day) for (a) CRU data, (b) HadAM3P, (c) PRECIS-Had and (d) PRECIS-ERA

42

Figure 3.11 Bias of annual precipitation (mm/day) for (a) PRECIS-Had and (b) PRECIS-ERA with respect to CRU data

44

Figure 3.12 Observed and simulated (PRECIS-Had and HadAM3P) annual cycle of precipitation averaged over the seven sub regions of figure 3.3

45

vii

Figure 3.13 Observed and simulated (PRECIS-ERA) annual cycle of precipitation averaged over the seven sub regions of figure 3.3

46

Figure 3.14 Observed (CRU) and simulated (PRECIS-Had and HadAM3P) seasonal precipitation coefficient of variation (CV) averaged over the seven sub regions of figure 3.3

47

Figure 3.15 Observed and simulated wet day frequencies averaged over the seven sub regions shown in figure 3.3.

48

Figure 3.16 Changes of mean temperature under SRES B2 scenario relative to present day climate (a) Annual, (b) Summer and (c) Winter

50

Figure 3.17 Changes of mean precipitation under SRES B2 scenario relative to present day climate (a) Annual, (b) Summer and (c) Winter

51

Figure 3.18 Annual cycle of temperature averaged over the seven sub regions for present (1961-1990) climate and future (2071-2100) climate under SRES B2 scenario

52

Figure 3.19 Annual cycle of precipitation averaged over the seven sub regions for present (1961-1990) climate and future (2071-2100) climate under SRES B2 scenario

53

Figure 4.1 Location map of Hunza, Gilgit and Astore river basins 58

Figure 4.2 Discharge of Hunza river, Gilgit river and Astore river during the rainfall events of (a) October, 1987 and (b) August, 1997

61

Figure 4.3 Mean annual cycle of temperature [˚C] over (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin as simulated with PRECIS RCMs and from CRU data

64

Figure 4.4 Mean annual cycle of precipitation [mm/day] over (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin as simulated with PRECIS RCMs and from CRU data

65

Figure 4.5 A schematic diagram of the hydrological model HBV (modified after Lindström et al., 1997), numbers in brackets refer to described equations

71

Figure 4.6 Sensitivity of HBV model parameters for (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin

77

Figure 4.7 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV-Met (Hunza river basin), with Y as a function of FC, GMELT, TT and DTTM

78

Figure 4.8 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV-ERA (Hunza river basin), with Y as a function of FC, GMELT, TT and DTTM

79

Figure 4.9 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV-PRECIS (Hunza river basin), with Y as a function of FC, GMELT, TT and DTTM

80

viii

Figure 4.10 Observed and simulated discharge (m3/s) of (a) HBV-Met, (b) HBV-ERA and (c) HBV-PRECIS for Hunza river basins during calibration period

81

Figure 4.11 Double mass-curve analysis relating observed and simulated discharge (m3/s) of (a) HBV-Met, (b) HBV-ERA and (c) HBV-PRECIS for Hunza river basin during calibration period

82

Figure 4.12 Observed and simulated (HBV-Met, HBV-PRECIS and HBV-ERA) mean annual discharge (m3/s) cycle of (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin

83

Figure 4.13 Observed, HBV-Met simulated and HBV-PRECIS simulated annual maximum discharge as a function of return period for three river basins in the present day climate

87

Figure 4.14 Observed discharge (green line) and uncertainties in discharge (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Hunza river basin during the 1986 hydrological year

88

Figure 4.15 Observed discharge (green line) and uncertainties in discharge (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Gilgit river basin during the 1986 hydrological year

89

Figure 4.16 Observed discharge (green line) and uncertainties in discharge (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Astore river basin during the 1986 hydrological year

90

Figure 5.1 Mean annual cycle of temperature [˚C] over river basins (a) Hunza, (b) Gilgit and (c) Astore as simulated with PRECIS for present (1961-90) and future (2071-2100) climate under SRES B2 scenario

95

Figure 5.2 Mean annual cycle of precipitation [mm/day] over river basins (a) Hunza, (b) Gilgit and (c) Astore as simulated with PRECIS for present (1961-90) and future (2071-2100) climate under SRES B2 scenario

96

Figure 5.3 Annual discharge cycle simulated by HBV-Met for the present climate and future SRES B2 climate for three stages of glaciation for three river basins

100

Figure 5.4 Annual discharge cycle simulated by HBV-PRECIS for the present climate and future SRES B2 climate for three stages of glaciation for three river basins

101

Figure 5.5 HBV-PRECIS simulated annual maximum discharge as a function of return period for current and changed SRES B2 climate for three glacier stages for three river basins

104

Figure 5.6 HBV-Met simulated annual maximum discharge as a function of return period for current and changed SRES B2 climate for three glacier stages for three river basins

105

ix

ABBREVIATIONS

ARD Absolute Relative Deviation

CRU Climate Research Unit

DTR Diurnal Temperature Range

ECMWF European Centre for Medium-Range Weather Forecasts

ENSO El Nino Southern Oscillation

GCM Global Climate Model

GHG Greenhouse Gas

HBV Hydrologiska Byråns Vattenbalansavdelning

HKH Hindukush-Karakorum-Himalaya

IPCC Intergovernmental Panel on Climate Change

NAO North Atlantic Oscillation

NCEP National Centers for Environmental Predictions

NS Nash-Sutcliffe

MSLP Mean Sea Level Pressure

PMD Pakistan Meteorological Department

PRECIS Providing REgional Climates for Impact Studies

RCM Regional Climate Model

SMHI Swedish Meteorological and Hydrological Institute

SIHP Snow and Ice Hydrology Project

SRES Special Report on Emission Scenarios

SST Sea Surface Temperature

UIB Upper Indus Basin

WGMS World Glacier Monitoring Service

WMO World Meteorological Organization

CONTENTS

ABSTRACT i

ACKNOWLEDGEMENTS iii

LIST OF TABLES iv

LIST OF FIGURES vi

ABBREVIATIONS ix

CHAPTER 1 INTRODUCTION 1

1.1 Background 1

1.2 Introduction to the Study Area 2

1.3 Hydro Meteorology of the UIB 7

1.4 Climate Change Impact Assessment of Water Resources 8

1.5 Objectives of the study 9

1.6 Thesis Layout 10

CHAPTER 2 LITERATURE REVIEW 12

2.1 Background 12

2.2 Climate Change Impact on Water Resources 13

2.3 Scenarios in Climate Change Studies 14

2.4 Global Climate Models (GCMs) 15

2.5 Downscaling of GCMs 16

2.5.1 Statistical downscaling 17

2.5.2 Dynamical downscaling 17

2.6 Water Resource Modelling under Climate Change 19

2.7 Uncertainty in Hydrological Impact Modelling 20

CHAPTER 3 ANALYSES OF PRECIS RCM CLIMATE CHANGE SCENARIOS

22

3.1 Background 22

3.2 Description of the PRECIS RCM 22

3.3 Representation of Topography in PRECIS 23

3.4 Experimental Design 24

3.4.1 Domain size and resolution 24

3.4.2 Boundary conditions 24

3.5 Present Day Climate Simulation Capacity of PRECIS 28

3.5.1 Mean sea level pressure patterns 29

3.5.2 PRECIS temperature simulations 30

3.5.3 PRECIS precipitation simulations 40

3.5.4 PRECIS estimated wet day frequency 48

3.5 Climate Change Responses under SRES B2 Scenario for Period 2071-2100

48

3.6 Summary 54

CHAPTER 4 PRECIS SIMULATIONS AS INPUT TO HYDROLOGICAL MODELLING

57

4.1 Background 57

4.2 Influence of Temperature and Precipitation on Discharge 59

4.3 Present Day Climate Data Analysis 59

4.3.1 Temperature 62

4.3.2 Precipitation 62

4.3.3 Bias correction in PRECIS simulations 66

4.4 River Basin Modelling 66

4.4.1 Description of HBV model 67

4.4.2 Model experiments 70

4.4.3 Calibration and validation of HBV models 72

4.4.4 Representation of flood peaks 84

4.4.5 Robustness of HBV models 84

4.5 Summary 91

CHAPTER 5 CLIMATE CHANGE IMPACT ON WATER RESOURCES

93

5.1 Background 93

5.2 Change of Temperature and Precipitation in the Selected River Basins

93

5.3 Climate Change Signals Transfer from PRECIS RCM to HBV 97

5.3.1 Delta change approach 97

5.3.2 Scaling approach 98

5.4 Assessment of Water Resources under Climate Change 98

5.4.1 Simulation of annual discharge cycle 98

5.4.2 Future discharge peaks 102

5.5 Summary 106

CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK

108

6.1 Conclusions 108

6.2 Recommendations for Future Research Work 111

REFERENCES 112

ANNEXURE LIST OF PUBLICATIONS 132

Chapter 1 Introduction

1

CHAPTER 1

INTRODUCTION

1.1 Background

Pakistan is a developing country. Its estimated population is over 162.6 million and about 76

% of this population lives in the country side (PCO, 2008; P & D, 2001). Its economy is

agro-based and highly dependent on the large scale Indus irrigation system (SIHP, 1990).

Forty million irrigated acres consume 100 million acre feet of water annually, which is

approximately 70 % of the total annual river runoff (WAPDA, 1990). Furthermore, per capita

water availability has been reportedly decreased from 5600 m3 to 1000 m3 since 1947

(Kahlown et al., 2007).

Climate change impact on the water resources is likely to affect irrigation system of Pakistan

(Wescoat, 1991). It has potential to affect the installed power capacity of the country as well.

Changes in flow magnitude in Indus river are likely to raise tensions among the provinces,

especially within the downstream areas because of reduced water flows in the dry season and

higher flows and resulting flood problems during the wet season. Changes in climate may

also increase the occurrence of hydrological extremes such as droughts and floods. This

situation demands to investigate the climate change impact on the present and future water

resources. It will provide a precise and comprehensive data base in order to conceptualize the

better strategies for water resource planning and management in terms of formulation of

policies for investments in irrigation system, agriculture, hydropower production and flood

protection measures.

In this study, an attempt has been made to assess the climate change impact on the future

river discharge in Pakistan. To achieve this end climate change scenarios are developed using

Providing Regional Climate for Impact Studies Regional Climate Model (PRECIS RCM).

The outputs from PRECIS RCM are used as input into the HBV (Hydrologiska Byråns

Vattenbalansavdelning) hydrological model to estimate the river discharge in the present and

future climate.


2

1.2 Introduction to the Study Area

The Indus River has a few large and many small tributaries (figure 1.1). The total area of

Indus river basin is about 970, 000 km2. However, this study is confined only to Upper Indus

Basin (UIB) that lies between the source of Indus river and Tarbela reservoir covering a

basin area of about 175, 000 km2 (NESPAK, 1997). Major tributaries of UIB region include

Shyok, Gilgit, Hunza and Astore rivers. Since the UIB constitutes a major part of Hindukush-

Karakorum-Himalaya (HKH) region, the terms UIB and HKH will now onward be used

alternatively in this study.

The water flow in streams of the study area is characterized by extreme seasonal variability

(figure 1.2). Some 80 % of total yield occurs in only 6-10 weeks of an average year. Seasonal

snowmelt and melting of glacier ice are both main contributors of river discharge, which is

obviously increased during summer because of higher temperature. Almost 80-90 % area of

UIB is snow covered with occasional exception of 60 % in winter season. The UIB lacks

major lakes and large forests and about one quarter of the area is occupied by glaciers

(SIHP,1990). The UIB rivers lie in the tectonically active regime (Seeber and Gornitz, 1983)

and have a potent danger of floods caused by landslides and rockslides (Hewitt, 1998).

Glaciers may also block the streams forming natural dams, which may cause floods on

rupturing (Hewitt, 1989).

The UIB region is dominated by large glaciers (table 1.1) and there is a fivefold to tenfold

increase in precipitation from glacier termini (~ 2500 m) to accumulation zones above 4800

m. Maximum precipitation occurs between 5000 and 6000 m (Hewitt et al. 1993). Most of

the glaciers are nourished mainly by avalanche snow. Westerly circulations and cyclonic

storms contribute two third of high altitude snowfall (Hewitt, et al. 1989), while one third

derives from summer snowfall mainly due to monsoon circulation (Wake, 1989). A huge loss

of ice mass and glacier recessions are observed in almost all Karakorum glaciers in 20th

century until the mid 1990s. Since then there has been thickening and advances in many

glaciers but confined to the highest watersheds of the central Karakorum (Hewitt, 2005). In

spite of surge type behavior of some glaciers in the HKH region (Diolaiuti et al., 2003), some

others (e.g. the Baltoro glacier) are stable during the last 100 years (Mayer et al. 2006) but

glaciers located in valleys are declining. According to weather stations records, there is a


3

shift towards positive mass balance thereby a reduction in run-off in the most heavily

glacierized Hunza basin (Fowler and Archer, 2006; Archer and Fowler, 2004). However,

sudden changes in glaciers and their confinement to the highest watersheds suggest that

thermal and hydrological threshold of glaciers is crossed, which triggers down slope

redistribution of ice by normal as well as surging flow with or without mass balance changes

(Hewitt, 2007).

In UIB, climatic variables are strongly influenced by altitude and an increase in snow pack

thickness is observed with height. Topography predominantly governs the spatial distribution

of snowfall and relative accumulation of snow and ice and their patterns of release by

melting. Heavy snowfall and accumulation of snow and ice occur at an altitude of above

3000 m. At an altitude of 2500 m, little precipitation and high evaporation is witnessed,

which may lead to severe aridity of the valleys. This phenomenon is more vivid at height

below 3000 m in the westerly and at an elevation of 4500 m in the easterly parts of the

Karakorum (SIHP, 1990). The HKH region receives a total annual rainfall in the range of

200-500 mm. Since these amounts are measured by valley-based stations and are not

representative for elevated zones. High-altitude precipitation estimates derived from

accumulation pits runoff above 4000 m range from 1000 mm to more than 3000 mm. These

estimates depend on the site and time of investigation as well as on the method applied

(Winger et al., 2005).


4

Figure 1.1 Indus River Basin


5

0

2000

4000

6000

8000

JAN MAR MAY JUL SEP NOVTime Period (Month)

Dis

char

ge (m

3 /s)

YogoAlam BridgeBeshameDoyianGilgitKachuraPartabShigarDainyorKarmongTarbela

Figure 1.2 Mean annual hydrograph at different locations at the Indus river including some headwater tributaries


6

Table 1.1 Dimensions of some large glaciers in the UIB region (Hewitt, 2003)

Elevations (m) Glacier

Length

(Km)

Area

(Km 2) Max Min Total Relief

Siachen 75.0 1181.0

Baltoro 62.1 756.3

Biafo 67.9 626.8

Hispar 53.1 621.6 7,885 3,040 4,845

Rimo 45.1 510.2

Skamri 41.0 427.4

Panmah 43.9 1515.1

Te Rong 27.4 295.3

Batura 59.6 285 7,795 2,540 5,255

Khurdopin 41 280 7,760 3,200 4,595

Sarpo Laggo 32.0 230.5

Braldu 35.1 202.0

Virjerab 36.1 189 6,600 3,600 3,000

Kero Lungma 20.9 150.2

Yazghil 30.2 145 7,880 3,055 4,825

Barpu 33.7 136 7,740 3,050 4,390

Malangutti 23 105 7,880 2,850 5,030

Yashkuk Y 24 125 6,700 3,500 2,900

Bualtar 21.5 105 7,200 2,450 4,850

Pasu 20.5 115 7,280 2,720 4,560

Ghulkin 18 55 7,310 2,420 4,890

Hassanabad 17 60 7,286 2,750 4,536

Minapin 16 58 7,196 2,350 4,846


7

1.3 Hydro Meteorology of the UIB

The climatology of the UIB is influenced by western disturbances and most of the

precipitation falls in winter and spring (Archer and Fowler, 2004; Treydte et al., 2006).

Indian monsoon systems bring occasional rain in UIB (Wake, 1987). Precipitation is found to

be increased during the late nineteenth and the twentieth centuries to yield the wettest

conditions of the past 1000 years (Treydte et al., 2006). Several researchers (Karl et al. 1993;

Easterling et al., 1997; Jones et al., 1999) have reported that globally mean temperature is

increasing and diurnal temperature range (DTR) is decreasing. However, things are other

way round in the UIB, especially over the western Himalayan region (Yadav et al., 2004;

Fowler and Archer, 2006; Klein-Tank et al., 2006; Bhutiyani et al., 2007). Where an increase

in DTR (Akhtar et al., 2005a; Fowler and Archer, 2006; Bhutiyani et al., 2007) and a cooling

of mean temperature during pre-monsoon season is reported (Kumar and Hingane, 1988;

Akhtar and Hussain, 2001; Archer, 2003; Archer, 2004; Akhtar et al., 2005b). It could be

perhaps due to local forcing factors such as topography and land use etc. The significant

variations in summer temperatures indicate potentially important impacts on river runoff

(Archer, 2003; Fowler and Archer, 2006). In addition, temperature trend along with

precipitation would influence glacier mass balance. Hewitt (1998, 2005) reports the

widespread expansion of larger glaciers in the central Karakorum while most of the world’s

mountain glaciers are reportedly shrinking during the 20th century. And at a warming rate of

0.04 K per annum, without increase in precipitation, few glaciers would survive until 2100

(Haeberli et al., 1999; WGMS, 2002; Mastny, 2000; Shrestha and Shrestha, 2004; Paul et. al.,

2004). If the warming rate is 0.01 K per annum and with an increase in precipitation of 10%,

it is predicted that overall loss would be restricted to 10 to 20% of the 1990 volume of the

glacier (Oerlemans et. al., 1998).

Indian summer monsoon is one of the most important components of the coupled ocean-land-

atmosphere system. Significant links have been identified between Indian monsoon rainfall

and global climate, including El Nino Southern Oscillation (ENSO) (Syed et al., 2006) and

Mediterranean pressure indices (Raicich et al., 2003). Both North Atlantic Oscillation (NAO)

and ENSO affect climate of UIB where a positive precipitation anomaly is found during the

positive NAO phase and warm ENSO conditions. Positive precipitation anomaly could also

be due to western disturbances as they are intensified over the region when are encountered


8

with low-pressure trough (Syed et al., 2006). There also exists an inverse relationship

between Indian monsoon rainfall and Euroasian winter snow cover (Bamzai and Shukla,

1999).

In the UIB, summer runoff is highly correlated with winter precipitation in middle-elevation

basins e.g. Gilgit and Astore river basins (Fowler and Archer, 2005). In contrast, summer

runoff in high-elevation river basins such as the Hunza and Shyok, fed by glaciers and

permanent snow pack, is not correlated with winter precipitation but highly linked with

summer temperature (Archer, 2003). Fowler and Archer (2006) estimated a 20 % decrease in

runoff in the Shyok and Hunza river basins resulted from 1°C fall in observed summer

temperature. Linear regression analysis by Archer (2003) suggests that a 1°C rise in mean

summer temperature arising from climate change would result in an increase of 17% and 16

% in summer runoff for the rivers Shyok and Hunza, respectively.

1.4 Climate Change Impact Assessment of Water Resources

Much efforts have been focused on the evaluation of climate change modelling and assessing

the merits and demerits of different downscaling methods but little work has been done on

the application of these methods for the hydrological impact assessment. For instance, there

were only ten publications on this important issue in 2005 (Fowler et al., 2007a). The climate

change impact assessment of future water resources is undertaken using a Global Climate

Model (GCM) data as input in hydrological models (Watson et al., 1996). However, the

hydrological modelling requires a GCM data of fine spatial resolution. One way to obtain

this resolution is through statistical downscaling (Wilby et al., 1999). Numerous researchers

(Bergström et al., 2001; Pilling and Jones, 2002; Guo et al., 2002; Arnell, 2003; Booij, 2005)

have attempted statistical downscaling of different GCMs for determining hydrological

responses to climate change. An alternative approach is to use fine scaled GCM data through

dynamical downscaling (Hay et al., 2002; Hay and Clark, 2003). In this approach, a regional

climate model (RCM) uses GCM data as initial and lateral boundary conditions over a region

of interest. The fine resolution of a RCM (about 25-50 km) is more appropriate for resolving

the influence of small-scale features of topography and land use on climatological variables.

Moreover, the high resolution of the RCM is ideal to capture the variability of precipitation

as input to hydrological models (Gutowski et al., 2003). Recent applications of RCM data in


9

hydrological impact studies are presented by other workers (Kay et al., 2006a, b; Fowler et

al., 2007b; Graham et al., 2007a,b; Leander and Buishand, 2007; Akhtar et al., 2008a) as

well.

Different scenarios of the future meteorological conditions (temperature and precipitation)

are used as input to a hydrological model of a river basin in order to calculate the

corresponding discharges. However, the outputs from RCMs are subject to systematic biases.

Thus, direct use of outputs from RCM present day simulations into hydrological model

simulations typically leads to considerable deviations in river discharge from observations

(Fowler et al., 2007a). To transfer the signals of climate change from the results of RCMs

into hydrological model an interface is required. One way is to apply these changes through

simple transformation rules. This approach is referred as the delta change approach (Hay et

al., 2002). In the delta change approach, an expected mean temperature change is added to

the observed temperature record to obtain a future temperature time series and precipitation

is usually multiplied by a fraction. Another way to estimate the future water resources is by

using RCM outputs directly in the hydrological model (Fowler et al., 2007b; Graham et al.,

2007b; Leander and Buishand, 2007). Nevertheless, some bias corrections are incorporated in

the RCM outputs before their use in hydrological models. The direct use of RCM output is

advantageous because of holding the physical correlation between the downscaled

temperature and precipitation (Fowler et al., 2007a). One of applications of hydrological

models is to create runoff scenarios for different glaciation conditions. However,

hydrological models (both conceptual and physical) are still not able to deal with glacier

storage completely. Hence, holistic approaches to study and model glacier storage are of

major importance to fully integrate glaciers into the hydrological balance to be used for water

resources and river flow predictions at all time scales (Jansson et. al., 2003).

1.5 Objectives of the Study

The main objective of this study is to assess the climate change impact on the water resources

in UIB of Pakistan.

The following objectives ensue from this major goal:

1. To generate high resolution climate scenarios using PRECIS regional climate model.


10

2. To evaluate the ability of PRECIS to simulate present day climate (1961-90) and to

predict the change in temperature and precipitation for the time period 2071-2100

under SRES B2 scenario.

3. To calibrate/validate hydrological model using different input data sources and to

investigate the performance of hydrological model.

4. To determine how well historic daily distribution of annual and seasonal flows is

predicted using RCM data as input into hydrological model.

5. To assess the impact of climate change on the future discharges and future peak

flows.

1.6 Thesis Layout

This thesis is comprised of six chapters. Chapter 1 highlights the rational of the study, briefly

introduces study area, summarizes the climate change impact assessment of water resources

and sets out the objective of the study.

Chapter 2 reviews the pertinent literature, including topics like hydro climatic changes in

present and future climate, developments in the evaluation of global climate models, methods

to downscale GCM data and ways to use climate change scenarios in water resource impact

studies.

Chapter 3 details the generation of high-resolution climate data using PRECIS model. It

describes the PRECIS regional climate model and its ability to simulate present day climate.

Future climate under SRES B2 scenario is also presented.

Chapter 4 describes the application of hydrological models for climate change impact

studies. It gives the description of HBV hydrological model and details its calibration and

validation procedure. The robustness of hydrological model is analyzed and ranges of

uncertainties are examined. The behavior of flood peaks at different return periods in the

present day climate is also simulated.

Chapter 5 assesses the climate change impact on future water resources. This is accomplished

by applying the climate change scenario to the hydrological model. Future water resources

and flood peaks are estimated under SRES B2 scenario at different stages of deglaciation.


11

Finally, Chapter 6 presents the conclusions of the research and suggests recommendations for

future work.

Chapter 2 Literature Review

12

CHAPTER 2

LITERATURE REVIEW

2.1 Background

There is several fold increase in the atmospheric contents of greenhouse gases (GHGs) due to

rapid industrialization. For instance, CO2 content has enhanced from preindustrial value of

about 280 ppm to 379 ppm in 2005 whereas concentration of CH4 has elevated from 715 ppb

to 1774 ppb during the same period (Forster et al., 2007). The higher contents of GHGs bring

a change in the radiative balance of the earth resulting into climate change in terms of

increase in temperature, change in precipitation pattern and probably a rise in the frequencies

of extreme events (IPCC, 2001; Meehl et al., 2007; Tett, et al., 2007). The unprecedented

warming in the past two decades of twentieth century, especially during the period spanned

over 1995–2006, is believed to be due to the anthropogenic forcing of climate (Mann and

Jones, 2003; Thorne et al., 2003; Trenberth et al., 2007). Meteorological data of the previous

century also suggest a global mean temperature rise of 0.07°C per decade (Folland et al.,

2001; Jones and Moberg, 2003).

Globally observed annual precipitation has reportedly increased ~ 0.98 % per decade in the

twentieth century (New et al., 2001). The intensity of extreme events has also increased

worldwide in this century (Sillmann and Roeckner, 2007). The global mean sea level has

risen by 10 to 20 cm. There has been a 40 % decline in Artic Sea ice thickness in late

summer to early autumn in the past 50 years (Kunkel et al., 1999; IPCC, 2001). The

frequency of severe floods in large river basins has increased during the 20th century (Milly

et al., 2002). The average annual discharge of fresh water from six of the largest Eurasian

rivers has increased by 7 % from 1936 to 1999 (Peterson et al., 2002).

Hence, majority of the scientific community now believes that climate change is certain

(IPCC, 2007; Saier, 2007). This chapter reviews the pertinent literature, including topics like

hydro climatic changes in present and future climate, developments in the evaluation of

global climate models, methods to downscale GCM data and ways to use climate change

scenarios in water resource impact studies.


13

2.2 Climate Change Impact on Water Resources

The average global surface temperature is projected to increase by 1.4-3 ˚C from 1990 to

2100 for low emission scenarios and 2.5-5.8 ˚C for higher emission scenarios of GHGs in the

atmosphere (IPCC, 2007). It is argued that warming escalates the moisture holding capacity

of the atmosphere, alters the hydrological cycle and changes the characteristics of

precipitation (Fowler and Hennessy, 1995; Trenberth, et al., 2003). Changes in the

precipitation may however have a greater impact on human well being and ecosystem

dynamics than the temperature (IPCC, 2001; Trenberth and Shea, 2005) because

precipitation controls the volume of runoff whereas changes in temperature mostly affect the

timing of runoff (Barnett et al., 2004). In a changed climate, runoff is expected to be

decreased in some parts of the world, including the Mediterranean regions, parts of Europe,

central and southern America, and southern Africa. In other parts of the world, particularly in

southern and eastern Asia, there is likelihood that runoff would increase under climate

change, but this increase in discharge may not be very beneficial because it tends to come

during the wet season and the extra water may not be available during the dry season (Arnell,

2004). Therefore, modifications of the climate can change hydrological regime, which may

affect hydropower production, irrigated agriculture and increase water related risks such as

flood and droughts (Willis and Bonvin, 1995; Loukas et al., 2002; Jasper et al., 2004).

Changes in climate may have an impact on water resource availability, an increase in the

frequency and intensity of floods, drought and low flows (Milly et al., 2002; Huntigton,

2006; IPCC, 2007). Consequently, a warmer and more dynamic climate may lead to the

intensification of the hydrological cycle (Fowler and Hennessy, 1995, Trenberth, 1998,

1999). For instance, in the Illecillewaet river basin of British Columbia, Canada, the

magnitude and temporal distribution of flood frequency will be decreased due to a decrease

of snow pack and earlier snowmelt (Loukas et. al., 2004). During low flow period, drier

summer will lead to a decrease in the average discharge of Meuse river basin at Borgharen

gauging station in the Netherlands (De Wit et al., 2007). In case of Rhine river basin, there

appears higher winter discharge because of intensified snowmelt and increased winter

precipitation and decreased summer discharge due to reduced winter snow storage and an

increase of evapotranspiration (Middelkoop et al., 2001). The increase in temperature


14

accompanied by a reduced precipitation will lead to a decrease in the discharge of Mulde

river basin, Southern Elbe, Germany (Menzel and Burger, 2002). The effect of climate

change on snow water equivalent, snowmelt runoff, glacial melt runoff and total stream flow

is examined for Spiti river, which is high altitude river located in the western Himalayan

region. It is found that with the change in temperature (1-3 ˚C) the annual snowmelt runoff,

glacial melt runoff and total stream flow has increased. However, the most prominent effect

of increase in temperature has been noticed on glacier melt runoff (Singh and Kumar, 1997).

This shows that climate change impacts on hydrological systems strongly depend on the

characteristics of studied hydro-climatic region (Mohseni and Stefan, 2001; Singh and

Bengtsson, 2005). For instance, hydrological regime of mountain areas is strongly influenced

by the water accumulation (in the form of snow and ice) and by the melt processes

(Middelkoop et al., 2001). An absence of melt water affects surface hydrology in

extratropical region causing significantly drier upper layer soils and changes the annual cycle

of runoff (Vavrus, 2007).

2.3 Scenarios in Climate Change Studies

Scenarios of future climate change are required to assess the impact of climate change on

various sectors such as water resources, food production and ecosystem. A climate scenario

is a plausible, self-consistent outcome of the future climate that has been constructed for

explicit use in investigating the potential consequences of anthropogenic climate projections.

These climate projections depend on the future changes in emissions or concentrations of

GHGs and other pollutants (e.g. sulphur dioxide), which in turn are based on assumptions

related to future socioeconomic and technological developments and are therefore subject to

substantial uncertainty. A first set of emission scenarios known as the IS92 scenarios is

developed by the IPCC in 1992 (Leggett et al., 1992), updated by the Special Report on

Emission Scenarios (SRES) (IPCC, 2000). SRES include six scenario groups like A1B, A2,

B1, B2, A1T and A1F1 (Nakicenovic et al., 2000). However, SRES A2 and B2 scenario are

widely used in climate change studies. Theses scenarios cover a range of approximately 60 %

of the full span of emission scenarios (Woth, 2006).

The A2 scenario family describes a heterogeneous world characterized as slow economic

growth and rapid population growth rate as compared to A1 scenario. The underlying theme


15

is self-reliance and preservation of local identities. Economic growth is regionally oriented,

and hence both income growth and technological change are regionally diverse. The B2

scenario family describe a world in which the emphasis is on local solutions to economic,

social and environmental sustainability. Population increases at a lower rate than A2 but at a

higher rate than A1 and B1. This scenario is oriented towards environmental protection and

social equity (Nakicenovic et al., 2000).

2.4 Global Climate Models (GCMs)

Although there are a number of approaches to construct global climate scenarios, GCMs are

believed to be the most sophisticated tools being used to simulate global scale climate (Perks

et al., 2000). Global climate models are based on fundamental laws of atmospheric physics

and manifest the earth’s atmosphere in three-dimensional mathematical representations.

GCMs describe physical relationships such as the processes governing cloud, precipitation

and radiation and are used to provide information on how the climate may evolve or change

under certain conditions. The equations that govern GCM operation describe changes in

momentum, temperature, moisture and subdivide the atmosphere vertically into discrete

layers (Raisanen et al., 2007). Examples of GCMs are CCSR/NIES developed in

collaboration between the University of Tokyo and National Institute of Environmental

Studies, Japan (Emori et al., 1999), CGCM1/2 developed by Canadian Centre for Climate

Modelling and Analysis (Flato et al., 2000), ECHAM4/OPYC3 is developed in co-operation

between the Max-Planck-Institute for Meteorology and Deutsches Klimarechenzentrum in

Hamburg, Germany (Legutke et al., 1999), HadCM2/3 developed at the Hadley Centre, UK

(Johns et al., 1997; Gordon et al., 2000) and ModelE of NASA, USA (Schmidt, 2006).

GCMs are used to conduct two types of experiments such as equilibrium and transient

experiments for the estimation of future climate (Loaiciga, 1996). Historically, most GCM

simulations have been made in equilibrium mode providing a scenario for a point in the

future where the climate system has reached a balance with a given increase in the

concentration of GHGs. The conditions considered in these experiments represent the

combined effects of all the GHGs that would be equivalent to the radiative forcing of a

double concentration of atmospheric CO2. However, in transient simulations of the climate

system, CO2 level is increased at a fixed rate. Transient experiments are generally performed


16

by using high resolution fully coupled global ocean atmospheric GCMs (Joubert and

Tyson,1996). It is known that GHGs cause warming whereas sulphate aerosols may produce

cooling (Piltz, 1998). Therefore, there is a need to differentiate GCM simulations including

sulphate forcing and those excluding sulphate aerosols. GCMs incorporating both greenhouse

gas and sulphate aerosol forcing give a better representation of the observed pattern of

temperature change (Santer et al., 1994; 1995; Santer, 1996). For instance, Stott et al., (2000)

show that their model simulates the 20th century global mean surface air temperature

remarkably well and warming in the second half of the century is the result of anthropogenic

increase in GHGs concentration. Raisanen et al., (2007) reviewed the reliability of climate

models and found that simulated and observed climate data showed good agreement for

many basic variables. However, the ability of GCMs in simulating the frequency of extreme

events is poor (Kharin and Zwiers, 2000; Kysely, 2002).

2.5 Downscaling of GCMs

GCMs have coarse resolution and are unable to resolve fine scale features such as

topography, clouds and land use (Grotch and MacCracken, 1991; Downing et al., 2002).

Therefore, their suitability for climate change impact assessment on various natural and

managed systems at regional scale is questioned (von Storch et al., 1993; Ciret and Sellers,

1998; Hellström and Chen, 2003; Gaffin et al., 2004; Linderson et al., 2004). Hence, there is

a need to get situation specific information about climate to investigate the climate change

impacts (Li and Sailor, 2000; van Vuuren et al., 2007). As a result, considerable efforts have

been focused on the development of techniques like downscaling in order to bridge the gap

of GCMs prediction skills. The approaches employed to change GCM data to a finer scale

could be broadly classified into two categories, statistical downscaling and dynamical

downscaling (Tripathi et al., 2006). Statistical downscaling method establishes an empirical

relationship between GCM climate variable and local climate (Karl, et al., 1990), whereas in

dynamical downscaling Regional Climate Model (RCM) is embedded within a GCM (Giorgi,

1990; Giorgi & Mearns, 1991). Wood et al., (2004) state that the minimum standard of any

useful downscaling method for hydrological applications requires the observed conditions to

be reproducible.


17

2.5.1 Statistical downscaling

In this technique, regional or local climate information is derived by first determining a

statistical model, which relates large-scale climate variables to regional and local variables.

Then the large-scale output of a GCM simulation is fed into this statistical model to estimate

the corresponding local and regional climate characteristics. Many statistical downscaling

techniques (Karl, et al., 1990; Burger, 1996; Linderson et al., 2004) have been developed to

translate large scale GCM output into a fine scale. The simplest scheme is perturbation

method or delta change approach (Prudhomme et al., 2002; Fowler et al., 2007a). In this

method, difference between the present and future GCM simulations is applied to baseline

observations by simply adding or scaling the mean climate change factor. Therefore, it can be

rapidly applied to several GCMs to produce a range of climate scenarios. The main

limitations of this method are that for precipitation the temporal sequence of wet days are

unchanged and change factor only scale the mean, maxima and minima of climatic variable.

Moreover, delta change method ignores change in variability and assumes spatial patterns of

climate to be constant (Diaz-Nieto and Wilby, 2005). More sophisticated statistical

downscaling methods can be divided into three main categories including weather generators,

regression models and weather classification schemes (Wilby et al., 2004). The primary

advantage of these techniques is that they are computationally inexpensive, and thus can be

easily applied to output from different GCM experiments. The main disadvantages of these

methods include a requirement of long and reliable observed historical data, dependence on

the selection of predictors and absence of climate system feedback (Wilby and Wigley, 1997;

Fowler et al., 2007a; Wetterhall et al., 2007).

2.5.2 Dynamical downscaling

This approach allows direct modelling of the dynamics of the physical system that

characterizes the climate of the region. Dynamical downscaling techniques can be grouped

into two classes such as high resolution and variable resolution atmospheric GCM (AGCM)

and Regional climate models (RCMs) (Jones et al., 2004). An AGCM is run for a specific

period of interest with boundary conditions of surface temperature and ice concentration. A

typical resolution of an AGCM is 100 km (May and Roeckner, 2001) whereas regional

experiments are performed at 50 km resolution (Deque et al., 1998). In AGCMs, atmospheric


18

and land-surface conditions interpolated from the corresponding GCM fields are used to

initialize the simulations. The main advantage of AGCM technique is that the resulting

simulations are globally consistent. However, the major disadvantage of this method is that it

is computationally very demanding (Jones et al., 2004).

RCM technique uses GCM data as lateral and initial conditions for selected time periods

(Dickinson et al., 1989; Giorgi, 1990). Information about other factors like sea surface

temperature, sea ice, greenhouse gas and aerosol forcing, as well as initial soil conditions are

also gathered from GCM data. The spatial patterns of climate produced by the RCMs are

usually in better agreement with observations compared to those of the GCMs. RCMs are

able to realistically simulate regional climate features such as orographic precipitation (Frei

et al., 2003), extreme climate events (Fowler et al., 2005; Frei et al., 2006) and regional scale

climate anomalies (Leung et al., 2003). However, model skill depends strongly on biases

inherited from the driving GCM and the presence and strength of regional scale forcing such

as orography, land-sea contrast and vegetation cover. Studies over regions where topographic

effects on temperature and precipitation are more prominent often reports more skilful RCM

downscaling than in areas where regional forcings are weak (Wang et al., 2004).

The main advantages of RCM include its ability to simulate high resolution information on a

large physically consistent set of climate variables and its better representation of extreme

events (Huntingford et al., 2002; Frei et al., 2003; Christensen and Christensen, 2003; Leung

et al., 2004). However, RCMs are computationally intensive and limited number of scenario

ensembles is available which restrict the model integrations to 30 years for present climate

from 1961-1990 and for a changed climate from 2071-2100 (Fowler et al., 2007a). This

makes climate change impacts for other periods difficult to assess. In RCMs, different

sources of uncertainty vary according to spatial domain, region and season (Deque et al.,

2005). However, the major contributor of uncertainties includes the errors in the driving

GCMs, RCM formulation and emission scenarios (Noguer et al., 1998; Christensen and

Christensen, 2001). Therefore, for each application careful consideration needs to be given to

some aspects of model configuration, such as physics parameterizations, model domain size

and resolution, and the technique for the assimilation of large scale meteorological forcing

(Giorgi and Mearns, 1999).


19

2.6 Water Resource Modelling under Climate Change

The most widely used approach to simulate the hydrological impacts of climate change is to

combine the output of GCMs with a deterministic hydrological model that contains

physically based or conceptual mathematical descriptions of hydrological phenomena. In

other words, the hydrological impacts of climate change on a watershed are investigated by

developing hydrological models of the watershed and simulating river flows that result from

total precipitation and temperature data derived from GCM outputs corresponding to specific

climate change scenario. A number of investigations have been conducted in this area during

the past few years. Loukas et al. (2002) investigated the potential impact of future climate

change on the causes of flood flows in different watersheds in British Columbia using the

UBC Watershed Model. Eckhardt and Ulbrich (2003) explored the impact of climate change

on groundwater recharge and stream flow in a central European catchment using a

conceptual echo-hydrological model (SWAT-G). Rosenberg et al. (2003) analysed the impact

of HadCM2 projections in 18 major water resource regions in USA using the SWAT

watershed model. Fowler et al., (2007b) examined the impacts of climate change on water

resources in north-west England by using the Mospa model, which is a complex water

resource planning model. Graham et al., (2007b) studied the hydrological impacts from

future climate change over Europe using two hydrological models including a conceptual

HBV model and physical based distributed WASIM model. Akhtar et al., (2008a) also used

the HBV model to investigate the impacts of climate change on the water resources of HKH

region under different glaciation scenarios.

A mismatch exists between climate and hydrologic modelling in terms of the spatial and

temporal scales, and between GCM accuracy and the hydrological importance of the

variables. In particular, the reproduction of observed spatial patterns of precipitation and

daily precipitation variability is not sufficient (Salathe, 2003; Burger and Chen, 2005).

Moreover, quality of GCM outputs prevents their direct use for hydrological impact studies

(Prudhomme et al., 2002). However, linking downscaled data to hydrological models may

improve the results (Wilby et al.,1999; Fowler et al., 2007a). The simplest methods is to use

hypothetical climate change scenarios by modifying time series of meteorological variables

by change factor in accordance with GCM scenario results (Arnell and Reynard, 1996;

Boorman and Sefton, 1997). However, this method does not allow for change in temporal


20

variability and so recent studies have used more sophisticated methods including

dynamically downscaled output (Graham et al., 2007a,b), bias-corrected dynamically

downscaled output (Wood et al., 2004; Fowler and Kilsby, 2007; Fowler et al., 2007b) and

statistical downscaling approaches (Pilling and Jones, 2002; Jasper et al., 2004).

The performance of different downscaling methods for hydrological impact assessment has

been assessed. Some studies suggest that statistical downscaling methods perform better

compared to the dynamical methods while simulating the snowmelt runoff in river basins

located in Colorado and Nevada, U.S (Hay and Clark, 2003). In case of river Thames, UK,

where runoff is not snowmelt driven, statistical downscaling methods perform poorly as

mean dry spell length is underestimated (Diaz-Nieto and Wilby, 2005). Kleinn et al., (2005)

report that using the bias corrected RCM data the discharge regime of Rhine river basin is

well simulated. Comparisons of different statistical downscaling methods give significantly

different hydrological impacts for the same river basin (Dibike and Coulibaly, 2005). This

may be because of the fact that statistical downscaling does not preserve correlation between

different downscaled variables. Moreover, statistical downscaling of precipitation is to be

less successful due to intermittent nature of precipitation (Li and Sailor, 2000) whereas RCM

gives better representation of precipitation variability (Hellström and Chen, 2003).

Additionally, the direct use of RCM data or use of bias corrected RCM data in impact studies

preserves the physical correlation between precipitation and temperature (Wood et al., 2004;

Fowler and Kilsby, 2007; Fowler et al., 2007a; Graham et al., 2007a,b).

2.7 Uncertainty in Hydrological Impact Modelling

In climate change impact studies, the quantification of hydrological modelling uncertainties

is important to assess whether the system modification is induced by climate change or by

model errors. These uncertainties arise from climate modelling, hydrological modelling and

from methods used to link climate change and hydrological models (Fowler et al., 2007a).

The hydrological modelling uncertainties are caused by different sources, including errors in

the input data, errors in recorded observations of phenomena to be modelled, errors in the

model structure and uncertainty due to model parameters (Refsgaard and Storm, 1996; Xu

and Singh, 2004). The observational errors are linked to measurement methods and to the

type of input required. A model is a simple representation of a natural phenomenon and


21

would not be able to produce outputs that match observations perfectly. This source of

modelling uncertainty is referred as the model structure uncertainty. The type of errors

induced by the parameters depends on how they are estimated. However, the best parameter

set is difficult to find and different parameter sets can yield equally good results for the

model calibration (Beven and Binley, 1992; Gupta et al., 1998). The climate modelling

uncertainties mainly contain systematic errors in GCM data and deficiencies in downscaling

approaches (Fowler et al., 2007a).

Booij, (2002) reports that the overall uncertainties in discharge due to input data errors are

more significant than uncertainties due to hydrological model errors and parameter

estimation errors. Wilby and Harris (2006) assessed the uncertainties in climate change

impacts on low flows in the river Thames, UK. They explored that the cumulative

distribution functions of low flows were most sensitive to uncertainties in downscaling of

different GCMs whereas uncertainties due to hydrological parameter estimation were found

to be less important. Hingray et al., (2007) concluded that the uncertainty resulting from

climate model is bigger than uncertainty introduced by the hydrological model. Krysonva et

al., (2007) found that the water discharge and the groundwater recharge in the Elbe basin will

be most likely decreased under expected climate change, but the uncertainty in hydrological

response to changing climate is generally higher than the uncertainty in climate input. Akhtar

et al., (2008b) studied the influence of different input forcing data on the discharge behavior

of river basins in HKH region. They found that the uncertainties were higher in the river

basins where discharge was dependant on precipitation compared to the river basins where

discharge was governed by temperature. In high mountainous areas, the available

meteorological and hydrological data are scarce and due to the extreme weather conditions

data are highly error prone. This problem represents a considerable source of uncertainty for

runoff and water balance simulation, especially in the presence of glaciers (Schafli, 2005).

For instance, in HKH region the future discharge estimated by using the meteorological

observations data is significantly different from the discharge predicted through RCM output,

mainly due to the errors in observed data (Akhtar et al., 2008a).

Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

22

CHAPTER 3

ANALYSES OF PRECIS RCM CLIMATE CHANGE SCENARIOS

3.1 Background

Hydrological modelling predominantly depends on the atmospheric parameters such as

precipitation, temperature and evapotranspiration. Precipitation in the form of snow and rain

is the major source of runoff generation and temperature influences the snowmelt processes.

Hence, reliable information of these parameters is an important prerequisite of the

hydrological modelling. The necessary information can be derived from observational

records. However, difficult topography of the area some times makes it inaccessible for

routine meteorological and climatological observations leading to scarcity of atmospheric

data. Therefore, for large-scale hydrological applications, especially for climate change

impact studies, RCM simulations can bridge this gap. RCMs generated regional data can

make it possible to drive other regional specific models analyzing local scale impacts

(Wilson et al., 2005). This study uses PRECIS (Providing REgional Climate for Impact

Studies) RCM to generate fine scale climate change scenarios as PRECIS has been used to

simulate features of present day climate in our region, including India and China (Kumar et

al., 2006; Yinlong et al., 2006; Lijuan et al., 2007).

3.2 Description of the PRECIS RCM

PRECIS is a high resolution atmospheric and land surface model, which covers a limited area

locatable over any part of the globe. It accounts for entities like dynamical flow, atmospheric

sulphur cycle, cloud and precipitation, radiative processes, land surface and the deep soil

coupled with the demarcation of boundary conditions (Jones et al., 2004). PRECIS is based

on the atmospheric component of the HadCM3 climate model (Gordon et al., 2000). The

atmospheric dynamics module of PRECIS is a hydrostatic version of the full primitive

equations and uses horizontal and vertical coordinates. There are 19 vertical levels, the

lowest at 825 hPa and the highest at 0.5 hPa (Cullen, 1993) with terrain following σ-

coordinates. An Arakawa B grid is used for horizontal discretization (Arakawa and Lamb,

1977) and horizontal diffusion is applied to control the accumulation of noise and energy at

the grid scale. PRECIS model also provides an option for the interactive reaction with sulfate


23

aerosols, which are simulated by using Lagrangian chemistry model STOCHEM (Collins et

al., 1997). The land-surface scheme employed in the PRECIS model is MOSES

(Meteorological Office Surface Exchange Scheme), which has shown good skill in land-

surface simulation (Bowling et al., 2003; Nijssen et al., 2003). It has a vegetated canopy that

provides fluxes of heat and moisture to the atmosphere and rainfall runoff. The MOSES uses

four soil layers in the vertical with depths chosen to capture important soil temperature

cycles. The scheme describes two components of runoff including surface runoff and

subsurface runoff (Cox et al., 1999 and Essery et al., 2003). The seasonal and daily varying

cycles of incoming solar radiation are also included. The boundary layers can occupy up to

the bottom five model layers. Observed sea surface temperatures (SSTs) and sea ice are used

for the base line climate simulations. For the future climate, changes in SSTs and sea ice

under SRES scenarios relative to baseline is derived from HadCM3 simulations (Jones et al.,

2004).

3.3 Representation of Topography in PRECIS

The representation of topography is an important input to climate models as it has a strong

impact on the simulated climate fields, especially in terms of spatial rainfall distribution. In

case of a vast flat terrain (thousands of kilometers) far away from coasts, the coarse

resolution of a GCM may not matter. However, HKH region has complex orographic

features, which play a key role in determining local climate and redistribution of solar energy

by surface interception. Precipitation is also highly sensitive to local topographic

characteristics and observations show that it increases with height (Winger et al., 2005).

Figure 3.1 shows the topographic features depicted by HadAM3P, PRECIS, GTOPO30

2MIN Digital Elevation Model (DEM) and the difference of surface elevation in HadAM3P

and PRECIS. The representation of elevation over HKH region is 500-3000 m higher in

PRECIS compared to the HadAM3P. To perceive the realism in the topographic details,

PRECIS and HadAM3P elevations are compared with the best estimates of topography i.e.

GTOPO30 2MIN DEM data. The representation of topographic features in PRECIS is quite

similar to that of GTOPO30 2MIN DEM (figure 3.1 b and c). However, there exist some

differences in orography, especially over the southwestern Pakistan where PRECIS results

show higher elevations compared to the GTOPO30. Anyhow, resolution of topographic


24

feature appears to be much better and PRECIS generated data are fit for the climate change

studies in the HKH region.

3.4 Experimental Design

Three experiments were designed in this study. Two experiments were aimed to simulate

present day climate and one experiment was exclusively for future climate simulation. The

experiments were mainly governed by factors such as study area, objectives of the study and

the ultimate use of PRECIS results. Table 3.1 gives some of the salient features of PRECIS

experiment. The brief descriptions of experimental components such as boundary data,

domain size and resolution of experiment are given in the following sections.

3.4.1 Domain size and resolution

The domain size is bounded by latitude 12 to 41 ˚N and longitude 55 to 97 ˚E (figure 3.2).

The horizontal resolution is 0.44˚ x 0.44˚ in rotation coordinates (~ 50 km). At this

resolution, the domain covers 89 grids in the longitude and 88 grids in the latitude. This

domain is reasonably large and covers most of South Asian region including India, Pakistan,

Afghanistan and Tibetan Plateau. It allows full development of internal mesoscale circulation

(e.g. monsoon circulation) and includes relevant regional forcings.

3.4.2 Boundary conditions

For the present climate simulations, PRECIS is nested with two global data sets. The global

forcing data from the ERA-40 reanalyses and HadAM3P GCM are used. The boundary data

of HadAM3P is the update version of atmospheric component of Hadley Centre coupled

ocean-atmospheric GCM HadCM3 in coarse resolution 3.75˚ in longitude and 2.5˚ in

latitude. To provide high-resolution boundary to PRECIS, HadAM3P is rerun in the

resolution of 1.875˚ in longitude and 1.25˚ in latitude based on the simulation of HadCM3.

For the baseline climate, it covers the period 1960-1990 (Wilson et al., 2005). ERA-40 is a

re-analysis of meteorological observations produced by the European Centre for Medium-

Range Weather Forecast (ECMWF). It covers the period 1957-2002, has a spatial resolution

of 1.875˚ by 1.25˚ and is extensively described in Uppala et al. (2005).


25

Figure 3.1 Topography of selected domain (a) Topography of the global climate model

(HadAM3P), (b) Topography of the regional climate model (PRECIS), (c) Topography of the GTOPO30 2MIN DEM and (d) Deviation of PRECIS RCM topography from GCM topography

a)

b)

Continued …….


26

Figure 3.1 Continued …….

c)

d)


27

For each global forcing, i.e. ERA-40 and HadAM3P, experiment were carried out at a

horizontal resolution of 50 x 50 km, hereafter referred to as PRECIS-ERA and PRECIS-Had,

respectively. The time period of PRECIS-ERA simulation was 1975-2001 whereas the

simulation of PRECIS-Had covers the period from 1960 to 1990. The period of PRECIS-

ERA experiment was restricted to start from 1975 because first 12 years ERA-40 boundary

data was corrupted (David Hein, personal communication). For future climate experiment the

global forcing data termed as HadAM3P: SRES B2 was used, hereafter referred to as

PRECIS HadB2. All experiments were performed by interfacing the sulphur cycle with

PRECIS. The first year in each experiment was considered as a spin-up period and data for

that period was not used in any analysis. After post processing of each experiment, the data

were prepared for further analysis.

Figure 3.2 PRECIS RCM domain for experiments at 50 x 50 km resolution


28

Table 3.1 Description of PRECIS RCM experiment

Experiment Driving Fields Resolution (km)

Number of Grids

Time Period

PRECIS-Had HadAM3P-Baseline 50 89 x 88 1960-1990

PRECIS-ERA ERA-40 50 89 x 88 1975-2001

PRECIS-HadB2 HadAM3P:B2 50 89 x 88 2070-2100

3.5 Present Day Climate Simulation Capacity of PRECIS The fine resolution regional simulations generated by PRECIS have been analyzed in detail

in order to assess the ability of PRECIS to model the regional climatological features. Model

validation was undertaken by following procedure after Giorgi et al., (2004). To evaluate the

systematic errors, the biases in PRECIS simulation were examined. The PRECIS simulated

temperature and precipitation was compared to the Climate Research Unit (CRU) data sets,

which is a 0.5° latitude/longitude gridded dataset of monthly observations for the period

1901-2002 (Mitchell and Jones, 2005). To estimate the errors caused by boundary data, the

PRECIS simulations were compared with the HadAM3P data. To analyze the errors in

internal model physics of PRECIS, we compared the biases in PRECIS-Had and PRECIS-

ERA climatology. For detailed analysis, land points of PRECIS domain were divided into

seven sub regions as shown in figure 3.3. The basic measure of temperature interannual

variability is the temporal standard deviation (STDV) as given in equation 3.1

1)( 2

N

TTSTDV i i (3.1)

where N is the number of years, Ti is the initial temperature and T is the average

temperature.

For the precipitation, the coefficient of variation (CV) was determined by the equation 3.2,

where the standard deviation of precipitation ( PSTDV ) is normalized by the average

precipitation ( P ) because PSTDV is affected by the mean, so that the CV is more

independent measure of interannual variability.

PSTDVCV P (3.2)


29

The ability of PRECIS to simulate patterns of mean sea level pressure (MSLP) was

investigated by comparing PRECIS simulated MSLS with those of NCEP reanalyses, ERA-

40 reanalyses and HadAM3P simulations (Kalnay et al., 1996; Uppala et al., 2005; Wilson et

al., 2005).

Figure 3.3 Sub regions used for more detailed analysis of the PRECIS RCM fields. Region 1 (Afghanistan), Region 2 (Southern Pakistan and Rajasthan), Region 3 (Hindu Kush-Karakorum- western Himalaya), Region 4 (Central Pakistan and Northwestern India), Region 5 (Tibetan Plateau), Region 6 (Central Himalaya) and Region 7 (Central India)

3.5.1 Mean sea level pressure patterns

The mean sea level pressure (MSLP) during the onset phase (May and June) of the Indian

summer monsoon is mainly important in order to understand the biases in regional

simulations. Figure 3.4 displays the MSLP for the period 1981-1990 for May-June (MJ) as

depicted by NCEP, ERA-40, HadAM3P, PRECIS-Had and PRECIS-ERA. A comparison of

MSLP patterns in NCEP reanalyses and ERA-40 reanalyses shows some differences over the


30

region. The main difference appears in the representation of heat low region. The minimum

pressure represented by ERA-40 reanalyses (~ 999 hPa) prevails over the Pakistan and

northwest India (Heat low region) lobbing towards the east coast of India. Whereas minimum

pressure represented by NCEP reanalyses (~ 1000 hPa) exists over the southern Pakistan and

Rajasthan. A trough of low pressure prevails over Tibetan Plateau, which is quite evident and

looks extended westward in ERA-40 data compared to NCEP data. The MSLP simulated by

both HadAM3P and PRECIS-Had (~ 994 hPa), over the heat low region, is more intense

compared to NCEP reanalyses. Kripalani et al. (2007) found similar pressure patterns in the

HadCM3 and linked the intensification of pressure systems to the decline in winter and

spring snowfall. The intense pressure systems in HadAM3P and PRECIS-Had may increase

the moisture and heat transport in the region. The heat low area in both HadAM3P and

PRECIS-Had simulations as compared to NCEP reanalyses appears to be shifted 1-2˚

northward. PRECIS-ERA and PRECIS-Had simulated MSLP are close to MSLP in ERA-40

and HadAM3P respectively. This indicates somehow the large-scale consistency between

RCM and driving forcing data. Compared to NCEP reanalyses data, HadAM3P and PRECIS-

Had simulate a very intense trough over Tibetan Plateau. Although PRECIS seems to

improve these anomalies in MSLP over heat low region and over the Tibetan Plateau,

however, the improvements are not much significant.

3.5.2 PRECIS temperature simulations

Figure 3.5 compares the observed (CRU) and simulated annual temperature. Compared to the

CRU observations the PRECIS-Had and PRECIS-ERA simulated temperature is relatively

low in the Tibetan Plateau and in the HKH region. The relatively low temperature is

substantive in PRECIS-Had as compared to HadAM3P and can best be attributed to the

steeps slopes escarpment over the Western Ghats, Tibetan Plateau and HKH region.

Compared to CRU data both PRECIS-Had and PRECIS-ERA show a general cold bias over

Tibetan Plateau and in the HKH region (figure 3.6). The presence of these errors in both

experiments suggests the inherent limitation of the model. Some of these cold biases may be

related to the positive precipitation bias because excess precipitation may cause extensive

wet soils resulting into high latent heat fluxes and low sensible heat fluxes. As a result

surface cooling (Bonan, 1998).


31

Figure 3.4 Mean seal level pressure (MSLP) for the period 1981-1990 for May-June (MJ)

season in (a) NCEP reanalyses data, (b) ERA-40 Reanalyses data, (c) HadAM3P GCM, (d) PRECIS-Had and (e) PRECIS-ERA

a)

b)

c)

Continued …….


32


Quantitative estimates of the temperature biases in the PRECIS-Had and HadAM3P can be

obtained from figure 3.7, which presents CRU, PRECIS-Had and HadAM3P temperatures

averaged over the subregions of figure 3.3. The annual cycle of temperature simulated by

both models matches reasonably well with the observed variations over all regions. Annual

cycle of temperature simulated by PRECIS-ERA also follows the patterns of observed

variations over all regions (figure 3.8). Generally, the profile of the temperature cycle shows

warm bias during the pre-monsoon season (i.e. April-June), while for the other seasons it is

underestimated. In monsoon-dominated regions (Region 4, Region 6 and Region 7), an

abrupt fall in temperature is observed during monsoon period. Kumar et al., 2006, also

observed this characteristic of PRECIS simulation nested in HadCM3 in all Indian mean

d)

e)


33

temperature. Table 3.2 shows that the subregional biases in the HadAM3P and PRECIS-Had

have a similar seasonal distribution. On average, the magnitude of temperature biases in

PRECIS-ERA is somewhat less compared to biases in PRECIS-Had. However, the

magnitude of biases during summer is generally higher in PRECIS–ERA compared to biases

in PRECIS-Had during the summer season. In all seven regions, the magnitude of biases is

higher during winter compared to summer season. Generally, the biases appeared to be

highest in the regions of complex topographic features (Region 3 and Region 5) and are

found to be lowest in relatively plane areas (Region 2 and Region 7). The cold bias observed

over mountainous areas (Region 3 and Region 5) is a common feature of regional climate

simulations over different regions of the world (Giorgi et al. 2004; Solman et al., 2008). It is

reported that CRU data of elevated areas may be warm biased due to the predominance of

less elevated stations and thus the simulated temperature may be underestimated in these

areas (New et al, 2000; Giorgi et al., 2004). The similar patterns of biases in both PRECIS-

Had and HadAM3P simulations demonstrate that the regional model have inherent biases due

to driving forcing.

Figure 3.9 shows the standard deviation of CRU temperature data and PRECIS-Had and

HadAM3P temperature simulations averaged over seven regions of figure 3.3. In January-

February-March (JFM) the temperature standard deviation varies in the range of 0.59 over

the central India (Region7) to about 1.39 in Afghanistan (Region 1). In July-August-

September (JAS), standard deviation for CRU data is more homogeneous in all regions and

varies in a narrow range of 0.34 to 0.64. Compared to the CRU observations, both the

PRECIS-Had and HadAM3P tend to overestimate the interannual variability in all seasons

except for the JAS where standard deviation of CRU data and PRECIS-Had and HadAM3P

simulations are in close agreement.

Generally, the standard deviations in the PRECIS-Had and HadAM3P are close to each other

which reflects that PRECIS inherent a good proportion of its temperature interannual

variability from global forcing data. Giorgi (2002) analyzed the dependence of surface

climate interannual variability on spatial scales and found that the temperature and

precipitation interannual variability tends to increase at finer scale, most markedly in the

summer. This scale affect on the interannual temperature variability is not very significant in

our PRECIS simulations.


34

Figure 3.5 Observed and simulated (baseline) patterns of annual temperature (˚C) for (a)

CRU data, (b) HadAM3P, (c) PRECIS-Had and (d) PRECIS-ERA

a)

b)

Continued …….


35


c)

d)


36

Figure 3.6 Bias of annual temperature (˚C) for (a) PRECIS-Had and (b) PRECIS-ERA with

respect to CRU data

a)

b)


37

Figure 3.7 Observed and simulated (PRECIS-Had and HadAM3P) annual cycle of

temperature averaged over the seven sub regions of figure 3.3

-505

1015202530

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

Tem

pera

ture

(˚C)

CRU PRECIS-Had HadAM3P

Region 1

05

10152025303540


Month

Tem

pera

ture

(˚C)


Region 2

-20-15-10

-505

101520


Month

Tem

pera

ture

(˚C)


Region 3

0

5

10

15

20

25

30


Month

Tem

pera

ture

(˚C)


Region 4

-25-20-15-10

-505

10


Month

Tem

pera

ture

(˚C)


Region 5

0

5

10

15

20

25


Month

Tem

pera

ture

(˚C)


Region 6

05

10152025303540


Month

Tem

pera

ture

(˚C

)


Region 7

Chapter 3 Analyses ofP:RECIS RCM Climate Change Scenarios

I

Figure 3.8

t ~~r~ 20~ 15

: 10~ 5

0

.......................................------.

............................


Region1Month

! 4--CRU - . .. - . .PRECIS-ERAI'

Q' 40

I

f"

~ 30: 20

11: 1

""'''''''''''''''''''''''''---'''''---'' ---.............

, .


MonthRegion2

I 4--CRU . - .PRECIS-ERAI

20Q:;,. 10~

~ 0

~.101 .'.~-20

Jan Feb Mar Apr May Jun Jut Aug Sep Oct Nov Dec

Region3 Month

1 4-- CRU - . .. - . .PRECIS-ERAI

-,

40 ........--..........;;

~ 30

i 20"J 10

0

--..........--................--.--..--.


Region4Month

I 4--CRU. . .. . . .PRECIS-ER~I

10,......--....--....----..--.............

;> 5

~ 0,;; -5~ -10~-15.:-20

-25 1 :Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec I

Month

Region5

1 4-- CRU . . .. - . .PRECIS-ERAJ

~ .-.

;> ~:

I

."""'---"''''''

.

---''''''~'~':''

... ,"~ 15 '.. .~ 10 ""J 5

0


Moolth

Region6

1 4-- CRU . .-:-. :-. . PR-ECIS.ER~J I

.J

~ ::j

:.~.."":':'.';':'.':':"':"."'''''''''''''':''''~ :~ 20 .' .'.. '~ '. :

J1: l' :Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Region7

Month

1 4-- CRU - . .. - . .~RECIS-ERAI

Observed and simulated (PRECIS-ERA) annual cycle of temperature averagedover the seven sub regions of figure 3.3

38


39

Figure 3.9 Observed (CRU) and simulated (PRECIS-Had and HadAM3P) seasonal temperature standard deviation averaged over the seven subregions of figure 3.3

Table 3.2 Biases in mean temperature (˚C) as simulated with the PRECIS-Had, PRECIS-ERA and HadAM3P relative to CRU reference data for different seasons and seven sub regions of figure 3.3 (Summer= April-September, Winter = October-March)

Temperature (˚C)

PRECIS-Had HadAM3P PRECIS-ERA Region Winter Summer Annual Winter Summer Annual Winter Summer Annual

Region 1 -2.18 1.57 -0.31 -2.53 1.21 -0.66 -0.68 0.71 0.02

Region 2 -1.85 -0.03 -0.94 -1.99 0.01 -0.99 0.74 1.25 1.00

Region 3 -6.75 -2.84 -4.80 -7.04 -2.90 -4.97 -5.04 -3.56 -4.30

Region 4 -3.95 -0.38 -2.16 -4.09 -0.29 -2.19 -1.02 0.46 -0.28

Region 5 -3.99 -1.72 -2.85 -3.51 -1.45 -2.48 -2.95 -2.38 -2.66

Region 6 -3.98 -1.31 -2.64 -3.85 -1.42 -2.63 -1.81 -0.31 -1.06

Region 7 -1.64 -1.41 -1.52 -1.47 -1.09 -1.28 -2.50 2.07 -0.21

Average -3.48 -0.87 -2.18 -3.49 -0.85 -2.17 -1.89 -0.25 -1.07

0.00

0.50

1.00

1.50

R1 R2 R3 R4 R5 R6 R7

Subregions

STD

V (˚

C)

CRU PRECIS HadAM3P

JAS

0.00

0.50

1.00

1.50

R1 R2 R3 R4 R5 R6 R7

Subregions

STD

V (˚

C)

CRU PRECIS HadAM3PJFM

0.00

0.50

1.00

1.50

R1 R2 R3 R4 R5 R6 R7

Subregions

STD

V (˚

C)

CRU PRECIS HadAM3P

AMJ

0.00

0.50

1.00

1.50

R1 R2 R3 R4 R5 R6 R7

Subregions

STD

V (˚

C)

CRU PRECIS HadAM3POND


40

3.5.3 PRECIS precipitation simulations

Observed (CRU data) and HadAM3P, PRECIS-Had and PRECIS-ERA simulated annual

precipitation patterns are shown in figure 3.10. The spatial distribution of precipitation in the

PRECIS simulation is very similar to the CRU data, which illustrates that the baseline

simulations adequately represent the present day conditions. However, some quantitative

biases in the spatial patterns also exist and the maximum biases are present over HKH region

(figure 3.11). The common biases in both PRECIS-Had and PRECIS-ERA experiments

indicate that these errors are due to deficiencies in the internal model physics. However,

some of the biases may be related to the inadequate representation of land surface in PRECIS

because seasonal variations in surface albedo, roughness and leaf area index could have a

significant effect on the climate (Hudson and Jones, 2002). The model currently uses

vegetation distribution and soil properties based on the climatology of Wilson and

Henderson-Sellers (1985) which does not account for these factors.

Figure 3.12 shows CRU observed, PRECIS-Had and HadAM3P simulated annual cycle of

precipitation whereas the annual cycle of precipitation simulated through PRECIS-ERA is

shown in figure 3.13. The annual cycle of precipitation as simulated through PRECIS-Had,

PRECIS-ERA and HadAM3P agrees well with observed variations in all regions. Over

Afghanistan (Region 1) and Hindu Kush-Karakorum-western Himalaya (Region 3)

precipitation reaches to maximum during the winter season, while precipitation over other

regions (Region 2, Region 4, Region 5, Region 6 and Region 7) is maximum during summer

season. Generally, it appears that PRECIS-Had and HadAM3P simulations overestimate

precipitation when compared to the observational data. Similar patterns are noted in

PRECIS-ERA simulations in all regions except for the Region 7 wherein PRECIS-ERA

underestimate precipitation. There are systematic differences in precipitation simulated

through PRECIS-Had and PRECIS-ERA as well. A systematic difference is also noted

between the precipitation simulated by HadAM3P and PRECIS-Had wherein the PRECIS-

Had tends to produce greater precipitation than HadAM3P. This is may be due to various

reasons. For example, Giorgi and Marinucci (1996) showed that the simulation of

precipitation may be sensitive to model resolution regardless of the topographic forcing. In

their experiments, precipitation tends to increase at finer resolution and topographic forcing

is found to be further strengthening this effect. Kumar et al. (2006) reported that during the


41

onset phase of the monsoon season, a significant bias is present in the Indian mean

precipitation estimated through PRECIS nested with HadCM3 boundary data. They inferred

that some of the biases in PRECIS simulation are inherited from boundary data. Notable

differences in PRECIS simulated precipitation over the Tibetan Plateau is also reported in a

study carried out in China (Yinlong et al., 2006). Kriplani et al., (2007) analyzed 22 GCMs to

investigate the South Asian summer monsoon variability, and found an intensification of

summer monsoon because of the intensification of low pressure over the Indo-Gigantic plain

and the land-ocean pressure gradient during the onset phase of the summer monsoon.

Therefore, the increase in precipitation in PRECIS-Had over the Tibetan Plateau and Hindu

Kush-Karakorum- Himalayan region could also be attributed to the intensification of the heat

low over these areas during the onset phase of the monsoon. Table 3.3 presents the

precipitation biases in PRECIS-Had, PRECIS-ERA and HadAM3P simulations compared to

the CRU data. The models generally overestimate the precipitation in all regions. However,

PRECIS-ERA is an exception, which underestimates precipitation in Region 1, Region 2 and

Region 7. On average, the precipitation biases in PRECIS-ERA simulations are somewhat

less compared to PRECIS-Had simulations. In some cases, HadAM3P gives slightly better

matching with CRU observations than PRECIS-Had. Many other researchers (Giorgi et al.

2004; Solman et al., 2008) strongly endorse this finding as well. They argued that the

apparent improvement in the GCM precipitation data was not due to the better representation

of the regional circulation features but owe to the compensations of model errors in the

GCM.

The coefficient of variation for precipitation is shown in figure 3.14. The observed

coefficient of variation ranging between 0.1-0.5 shows a relatively uniform distribution

throughout the region. The interannual variability of precipitation is generally higher during

the winter compared to the summer. This is due to the effect of dividing the standard

deviation by low precipitation amounts. Compared to the CRU observations, both the

PRECIS-Had and HadAM3P shows an underestimation of interannual variability in the high

mountain regions (Region 3 and Region 5) and overestimation in relatively flat regions

(Region 2 and Region 7). The PRECIS-Had and HadAM3P coefficients of variations are

generally in good agreement which shows that PRECIS obtain a good proportion of its

precipitation interannual variability from boundary data.


42

Figure 3.10 Observed and simulated (baseline) patterns of annual precipitation (mm/day)

for (a) CRU data, (b) HadAM3P, (c) PRECIS-Had and (d) PRECIS-ERA

a)

b)

Continued …….


43


c)

d)


44

Figure 3.11 Bias of annual precipitation (mm/day) for (a) PRECIS-Had and (b) PRECIS-

ERA with respect to CRU data

a)

b)


45

Figure 3.12 Observed and simulated (PRECIS-Had and HadAM3P) annual cycle of precipitation averaged over the seven sub regions of figure 3.3

0

0.5

1

1.5

2

2.5


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-Had HadAM3PRegion 1

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 2

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 3

0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 4

0

1

2

3

4

5

6


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 5

0

2

4

6

8

10

12


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 6

02468

101214


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 7


46

Figure 3.13 Observed and simulated (PRECIS-ERA) annual cycle of precipitation averaged

over the seven sub regions of figure 3.3

0

0.5

1

1.5

2

2.5


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERARegion 1

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERA

Region 2

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERA

Region 3

012345678


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERA

Region 4

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERA

Region 5

0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERA

Region 6

0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)

CRU PRECIS-ERA

Region 7

i ~~ 1..~...,.:...i...~..i ;...'..~...J i...~....0.00 L---i--l-

Rl R2 R3 R4 R5 R6 R7

SubregionsJFM

~PRECIS .II>.HadAM3P I

100 1" ,"":"'"'''''';'''''''''''':''''''''''';'''''''''''':''''''''''~'''''''''''

# 0 751-_n _~_nnn:n_n_n:nnnn:---nn"

I. ::: l ~--i--:---i---~---:---:-_.:-_I- ; ~---:0.00 -- --;

R1 R2 R3 R4 R5 R6 R7

Subregions

AMJFCRU .PRECIS .II>.HadAM3P I

'"''1 00

I :::[~'''''''''.:'''''''''':,,-'';'' ",'" : -'. - - ,

" O.<O'.rO'.:m m i '0.. 0 ,.~ '

I::.oili'i:~~" +-;;_10.00 '

R2 R3 R4 R5Rl

Subregions

JAS

10 CRU . PRECIS.II>. H;;dAM 3P J

--,

f :.: l~iiIoI~:.:m~...~ I

~ 0 ;11:' .[; 025. n_n_;nn: - -: On. n, , .

, , '

0 .00 '---u_+ m~ +.._u '-"._"", ,,;

Rl R2 R3 R4 R5 R6 R7

Subregions

ON

10CRU . PRECIS .II>.HadA~

Observed (CRU) and simulated (PRECIS-Had and HadAM3P) seasonalprecipitation coefficient of variation (CV) averaged over the seven sub regionsof figure 3,3 '

Figure 3,14

47


Table 3,3 Biases in mean precipitation (%) as simulated with the PRECIS-Had, PRECIS-ERA and HadAM3P relative to CRU reference data for different seasons and

seven sub regions of figure 3,3 (Summer= April-September, Winter = October-t March)

Precipitation (%)

Region PRECIS-Had HadAM3P 0 PRECIS-ERA

Winter Summer Annual Winter Summer Annual Winter Summer Annual

Region I -15 58 22 -5 57 26 -15 -1 -8

Region 2 36 46 41 24 32 28 -45 -17 -31

Region 3 92 101 97 76 55 65 126 109 117

Region 4 132 77 105 102 25 63 62 24 43

Region 5 21 102 61 36 83 60 33 77 55

Region 6 99 50 75 78 30' 54 27 7 17

Region 7 6 65 36 ,.., 33 15 -59 -41 -50-.)

Average 53 71 62 44 45 44 18 23 21


48

3.4.4 PRECIS estimated wet day frequency

Besides the evaluation of mean precipitation, we also analyze the ability of regional model in

simulating high frequency precipitation statistics in terms of wet day frequency, which is the

number of days per season with precipitation amounting greater than 0.1 mm. Figure 3.15

shows the frequency of wet days. It is evident that there is a close agreement in the simulated

PRECIS-Had and HadAM3P frequency of wet days with CRU observed frequencies in

Afghanistan (Region 1) and in central Himalaya (Region 4). The simulated wet days are

underestimated in southern Pakistan and Rajasthan (Region 2) and in central India (Region

7). An exceptionally high overestimation of frequency of wet days appears over Hindu Kush-

Karakorum- western Himalaya (Region 3) and Tibetan Plateau (Region 5).This

overestimation could better be explained due to the presence of substantially high pressure

over these regions.

0.00

0.20

0.40

0.60

0.80

1.00

R1 R2 R3 R4 R5 R6 R7

Region

Wet

day

freq

uenc

y

CRU PRECIS HadAM3P

Figure 3.15 Observed and simulated wet day frequencies averaged over the seven sub regions shown in figure 3.3

3.5 Climate Change Responses under SRES B2 Scenario for Period 2071-2100

PRECIS simulated annual, summer and winter spatial patterns of temperature change over

the period 2071-2100 for SRES B2 scenarios is illustrated in figure 3.16. Annual, summer

and winter spatial pattern of temperature change indicates an overall warming and maximum

warming is predicted over the western Pakistan. However, warming signals are weak over

the Himalayan mountain range. Infact, the cooling is likely in some small patches over the


49

Jammu and Kashmir region. Generally, winter season is more warmer compared to the

summer season. The mean annual cycle of temperature for present and future climate,

averaged over the seven sub regions, is presented in figure 3.17. A future increase in

temperature is evident over all sub regions. The future temperature follows the cyclic pattern

of present day climate. The predicted change in temperature is given in table 3.4. The mean

annual temperature change ranges between 2.9-3.3 ˚C. The maximum temperature change is

expected over Region 2 where predicted winter temperature is 3.4 ˚C. In the Region 3 (study

area), mean annual temperature change is 3.1 ˚C. On average, the annual mean temperature

rise is 3.1 ˚C and winter is 0.2 ˚C warmer than summer. In monsoon-dominated region 7, the

difference in summer and winter temperature change (0.6 ˚C) is relatively large.

Annual, summer and winter spatial patterns of precipitation change during 2071-2100 as

simulated by the PRECIS under SRES B2 scenario is given in figure 3.18. An annual mean

precipitation appears to be increased over most areas with a maximum over the Himalayan

mountain range, western and eastern Ghats. Whereas low precipitation change is evident

predominantly over border areas between Pakistan and India i.e. eastern parts of Sindh

Pakistan, Rajasthan India and southern parts of Punjab (both Indian and Pakistani parts of

Punjab). The spatial patterns of summer and winter precipitation change are similar.

However, the magnitude of summer precipitation change is higher compared to winter

precipitation change. The mean annual cycles of precipitation for present and future

precipitation are presented in figure 3.19. The future annual precipitation cycle follows the

pattern of present day precipitation. This shows that major shift in seasons is not expected in

future. In Regions 4 and 7, precipitation is expected to increase during summer season

whereas in Regions 1 and 3 the increase is predicted during winter season. There is an overall

increase in precipitation in SRES B2 scenario. Table 3.4 presents the annual and seasonal

changes in precipitation under SRES B2 scenarios for seven sub regions. Out of which only

Region 1 (i.e. Afghanistan) has shown a 7 % decrease in the mean annual future

precipitation. The rise in mean annual precipitation ranges from 1 to 21 %. In monsoon

dominated regions (i.e. Regions 4, 6 and 7), the precipitation during the summer season is

predicted to be increased up to 15%. In Region 3 (covering most of the UIB), the winter

precipitation with an expected rise of 3 % is important from the hydrological point of view.


50

Figure 3.16 Changes of mean temperature under SRES B2 scenario relative to present day

climate (a) Annual, (b) Summer and (c) Winter

a)

b)

c)


51

Figure 3.17 Changes of mean precipitation under SRES B2 scenario relative to present day

climate (a) Annual, (b) Summer and (c) Winter

a)

b)

c)


52

Figure 3.18 Annual cycle of temperature averaged over the seven sub regions for present

(1961-1990) climate and future (2071-2100) climate under SRES B2 scenario

-10

0

10

20

30

40


Month

Tem

pera

ture

(˚C

)

Present Temperature Future Temperature

Region 1

-10

0

10

20

30

40


Month

Tem

pera

ture

(˚C)


Region 2

-30

-20

-10

0

10

20


Month

Tem

pera

ture

(˚C

)


Region 3

-10

0

10

20

30

40


Month

Tem

pera

ture

(˚C

)


Region 4

-30

-20

-10

0

10

20


Month

Tem

pera

ture

(˚C

)Present Temperature Future Temperature

Region 5

-10

0

10

20

30

40


Month

Tem

pera

ture

(˚C

)


Region 6

0

10

20

30

40


Month

Tem

pera

ture

(˚C

)


Region 7


53

Figure 3.19 Annual cycle of precipitation averaged over the seven sub regions for present

(1961-1990) climate and future (2071-2100) climate under SRES B2 scenario

0

1

2


Month

Pre

cipi

tatio

n (m

m/d

ay)

Present Precipitation Future Precipitation

Region 1

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 2

0

1

2

3

4

5


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 3

0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 4

0

2

4

6

8


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 5

0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 6

02468

10121416


Month

Pre

cipi

tatio

n (m

m/d

ay)


Region 7


54

Table 3.4 Seasonal changes of mean temperature and precipitation under SRES B2 scenario

from PRECIS in 2071-2100 over the seven sub regions relative to 1961-1990 (Summer = April-September; Winter=October-March)

Temperature Change (˚C) Precipitation Change (%) Region Annual Winter Summer Annual Winter Summer

Region 1 3.2 3.3 3.2 -7 -8 -7

Region 2 3.3 3.2 3.4 1 -6 9

Region 3 3.1 3.1 3.2 6 8 3

Region 4 3.1 3.0 3.3 10 7 12

Region 5 2.9 2.9 2.8 21 10 32

Region 6 3.1 2.9 3.3 9 1 18

Region 7 2.9 2.6 3.2 15 15 16

Average 3.1 3.0 3.2 8 4 12

3.6 Summary

To simulate regional climate scenarios we have used PRECIS regional climate modelling

system, which is successfully set up in the South Asian region in some recent studies (Kumar

et al., 2006; Yinlong et al., 2006; Islam et al., 2008; Akhtar et al., 2008a,b). The scenarios

presented here are very useful for the impact assessment in various sectors. The PRECIS

output contains a large number of parameters. However, this study has focused mainly on the

temperature and precipitation. The analysis of the simulation mainly evaluates the capability

of PRECIS in representing spatial patterns of seasonal mean climate, its annual cycle,

interannual variability, wet day frequency and future change in temperature and precipitation

under SRES B2 scenario.

PRECIS improves the representation of mean climate compared to boundary data in many

aspects. The first feature to note is the representation of heat low area. Nevertheless, it

produces the intense pressure system and fails to produce the correct position of the heat low


55

compared to the NCEP reanalyses even then the representation of mean climate is better

compared to boundary forcing data. Kripalani et al. (2007) studied 22 GCMs to investigate

the South Asian summer monsoon variability and found an intensification of the heat low

over northwest India during the beginning of monsoon season. They linked intensification of

pressure systems with the decline in winter and spring snowfall. The intensity of heat low in

HadCM3 and ECHAM5 models investigated by them is comparable to HadAM3P and

PRECIS-Had simulations in our study. PRECIS simulated MSLP are close to MSLP of

driving forcing data (as expected) and this indicates somehow the large-scale consistency

between RCM and driving GCM.

The mean spatial patterns of temperature and precipitation agree reasonably well with CRU

observations though some model biases have been identified. For precipitation, biases

include an overestimation of precipitation especially over the mountainous regions.

Overestimation of precipitation is a common behavior in regional simulations over elevated

terrain of different parts of the world (Leung et al., 2003; Giorgi et al., 2004; Solman et al.,

2008; Islam et al., 2008). Therefore, higher orography in the PRECIS compared to GCM may

produce more precipitation (and more rainy days) over the mountain areas. This increase

could also be attributed to an intensification of the heat low. Kriplani et al., (2007) also found

an increase in summer monsoon because of the escalation of low pressure over the heat low

region.

Temperature shows in general a warm bias during the pre-monsoon months (i.e. April-June)

whereas for remaining months it is underestimated. Mean annual temperature cycle shows an

abrupt fall during monsoon period, which is a characteristic of all Indian mean temperature

(Kumar et al., 2006). The magnitude of temperature and precipitation biases in PRECIS-ERA

is somewhat less compared to PRECIS-Had. Overall, the magnitude of cold bias in

temperature is higher during winter compared to summer season bias whereas the magnitude

of precipitation bias is higher during summer compared to winter season bias. The seasonal

variations in temperature biases may be due to the variations in the latent heat flux in

different seasons (Uchiyama et al., 2006) which may not be well distinguished by the

PRECIS as well as by the driving forcing. The biases are higher generally in the region with

complex topographic features (e.g. Regions 3 and 5) compared to relatively plane areas (e.g.

Regions 2 and 7). This is a common feature of regional climate simulations found in different


56

parts of the world as well (Giorgi et al. 2004; Solman et al., 2008). Some of the biases may

also be related to the errors in CRU data (New et al, 2000; Giorgi et al., 2004) .Generally, the

estimated biases are comparable with the biases found in other studies in South Asian

countries ( Kumar et al., 2006; Yinlong et al., 2006; Islam et al., 2008).

On average, the interannual variability of simulated temperature is slightly higher compared

to the interannual variability in observed temperature. The interannual variability of

precipitation is somewhat less compared to the interannual variability in observed

precipitation in high mountain regions (Regions 3 and 5) whereas it is overestimated in

relatively flat regions (Regions 2 and 7). The similar values of interannual variability of both

temperature and precipitation in both PRECIS-Had and HadAM3P depict that PRECIS

inheriting a good proportion of its interannual variability from HadAM3P. Giorgi (2002)

found that the temperature and precipitation interannual variability tends to increase at finer

scale. However, this scale effect is not found in our simulations which is may be because of

same model physics used in PRECIS and HadAM3P. Despite the systematic errors discussed

here, the results are encouraging since the dynamical downscaling techniques are the most

reliable tool to estimate future projections of climate change with enough spatial detail, as

needed for impact studies.

PRECIS RCM simulation under SRES B2 scenario shows marked increase in temperature

and precipitation by the end of 21st century relative to the present day climate. However, the

spatial patterns of mean temperature changes show that over the Himalayan region, the

warming signals are week. Considering the average of seven regions (roughly average of all

land points) the average annual increase in temperature is predicted to be 3.1 ˚C whereas

precipitation is expected to rise by 8 %.

Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

57

CHAPTER 4

PRECIS SIMULATIONS AS INPUT TO HYDROLOGICAL MODELLING

4.1 Background

Conceptual water balance models are often believed to be useful in assessing the impact of

climate change on the regional hydrology (Arnell, 1999). These models have several

advantages over lumped models and/or physically based models. Main advantages include

flexibility and ease of use (Xu, 1999; Booij, 2002; Te Liunde et al., 2007). The most

important climatological inputs required for the calibration and validation of hydrological

models are temperature and precipitation that can be derived from observational records or

alternatively from regional climate models (RCMs). Meteorological data has a considerable

influence on the performance of hydrological model. For instance, inadequate representation

of spatial variability of precipitation in modelling can be partially responsible for modelling

errors. This may also lead to problems in parameter estimation of conceptual hydrological

models. It is reported that the uncertainties in discharge due to errors in meteorological inputs

are larger than uncertainties in hydrological model errors and parameter estimation errors

(Booij, 2002; Te Liunde et al., 2007). Therefore, a better understanding of the use of

meteorological data from various sources (observations and RCM simulations) in

hydrological models will enhance the confidence in predicted hydrological change.

Hence, the effect of precipitation and temperature, simulated with PRECIS RCM nested in

different global data sets, on the discharge simulated with the HBV model is examined.

Three river basins including Hunza, Gilgit and Astore are chosen to study the effect of

meteorological data on simulated discharge. Figure 4.1 gives the location and table 4.1 enlists

the salient features of these river basins of the study area (UIB region). The basin areas of

Hunza, Gilgit and Astore rivers are about 13925 km2, 12800 km2 and 3750 km2, respectively

and 34 %, 7 % and 16 % of which are glacial covered, respectively.


58

Figure 4.1 Location map of Hunza, Gilgit and Astore river basins

Table 4.1 Characteristics of study area

River Basins Parameters Hunza Gilgit Astore

Gauging Station Dainyor Gilgit Doyian

Latitude 35˚ 56‘ 35˚ 56‘ 35˚ 33‘

Longitude 74˚ 23‘ 74˚ 18‘ 74˚ 42‘

Elevation of gauging station (m) 1450 1430 1583

Drainage area (km2) 13925 12800 3750

Glacier covered area (km2) 4688 915 612

Mean elevation (m) 4472 3740 3921

% area above 5000 meter 35.8 2.9 2.8

No. of meteorological stations

Precipitation - 2 1

Temperature - 2 1

No. of PRECIS grid points at 50 km 6 5 2


59

4.2 Influence of Temperature and Precipitation on Discharge

Understanding of hydrological regime in UIB requires the joint consideration of affect of

temperature during extreme rainfall events and their impact on runoff. To depict the role of

temperature, two extreme rainfall spells are examined (table 4.2). The storm event of 11

October 1987 appeared to be one of the most sever episodes and spread over the entire area

of northern Pakistan. Precipitation occurred from 10 to 13 October but with heavy falls on 11

October resulting in a drop in mean daily temperature. In case of storm event of 28 August

1997, precipitation occurred from 27 to 28 August with maximum rainfall on 28 August

resulting in a decline in mean daily temperature. In both cases, daily maximum temperature

is more affected as compared to daily minimum temperature. Figure 4.2 shows the discharge

pattern of Hunza, Gilgit and Astore rivers during these two events. Monsoon rainfall record

suggests that the direct contribution of rainfall to river flow in the highly glaciated Hunza

river basin is small and the occurrence of rainfall is accompanied by a sharp fall in flood

hydrograph. The reduction in melt runoff in high altitude basin (Hunza river) is generally due

to reduced temperature and energy input. However, the hydrograph of Gilgit and Astore

rivers shows a small rise in the discharge during these rainfall events. The discharge behavior

of these river basins shows that both the energy inputs and intensity of rainfall events are

important considerations for hydrological modelling in the UIB.

4.3 Present Day Climate Data Analysis

In this section, temporal patterns of observed and simulated present day climate over selected

river basins are examined. CRU observed, PRECIS-Had and PRECIS-ERA simulated

temperature and precipitation data are extracted for each river basin. These data sets are

compared with each other to determine the ability of simulated fields in representing the

climate of river basins. The biases between PRECIS simulated temperature and precipitation

data and CRU reference data for different seasons are also presented.


60

Table 4.2 Temperature and precipitation during two monsoon events at selected stations

Temperature (°C) Precipitation (mm) Ev

ent Station Date Max. Min. Mean Hr.12 of

previous date to Hr. 0 of date

Hr.0 To Hr.3 of Date

Past 24 Hrs. Preceding Hr. 3 Of

Date

Hr.3 To Hr.12

of Date

9 22 9 15.5 0.0 0.0 0.0 -1.0 10 19 2.2 10.6 33.0 0.0 33.0 14.2 11 4.2 0.3 2.25 5.0 0.0 19.2 4.7 12 8.4 2.2 5.3 6.5 0.0 11.2 23.8

Astore

13 6.4 0 3.2 5.0 0.0 28.8 2.3

9 31 10.8 20.9 0.0 0.0 0.0 0.0 10 26.7 11.9 19.3 6.0 3.0 9.0 2.5 11 13.9 7.8 10.85 42.0 10.0 54.5 2.1 12 9.8 7.8 8.8 0.4 0.0 2.5 8.9

Gilgit

13 10.8 6.7 8.75 1.3 0.0 10.2 -1.0

9 27.8 11 19.4 0.0 0.0 0.0 0.0 10 24.9 9.6 17.25 8.0 8.0 16.0 226.0 11 13.3 4.5 8.9 295.0 5.0 526.0 41.0 12 9.3 5.4 7.35 114.0 0.0 155.0 23.0

A

Skardu

13 12.8 5.6 9.2 46.0 13.0 82.0 11.0

26 25.6 11.7 18.65 0.0 0.0 0.0 0.0 27 22.8 6.7 14.75 32.0 6.8 38.8 39.6 28 8.1 5.6 6.85 12.2 0.0 51.8 0.0 29 13.9 7.8 10.85 0.0 0.0 0.0 0.0

Astore

30 20.1 10.1 15.1 0.0 -1.0 0.0 3.0

26 27.5 17 22.25 0.0 0.0 1.2 0.0 27 29.8 14.2 22 20.0 13.6 33.6 34.6 28 14.8 9.8 12.3 14.1 0.0 48.7 0.0 29 23 11.8 17.4 0.0 0.0 0.0 0.0

Gilgit

30 28 14.5 21.25 0.0 0.0 0.0 0.0

26 30 14.2 22.1 0.0 0.0 0.0 1.0 27 23.3 11.3 17.3 15.0 2.0 18.0 19.0 28 12.2 9 10.6 14.7 0.0 33.7 0.0 29 18.3 10.3 14.3 0.0 0.0 0.0 0.0

B

Skardu

30 25.3 13 19.15 0.0 0.0 0.0 0.0

A: Event of October, 1987, B: Event of August 1997


61

Figure 4.2 Discharge of Hunza river, Gilgit river and Astore river during the rainfall events of (a) October, 1987 and (b) August, 1997

0

400

800

1200

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Day

Disc

harg

e (m

3 /s)

Hunza Gilgit Astore

b) Discharge during August 20 to September 8, 1997

0

200

400

600

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Day

Dis

char

ge (m

3 /s)

Hunza Gilgit Astore

a) Discharge during October 1-20, 1987


62

4.3.1 Temperature

Figure 4.3 compares the simulated and observed mean annual temperature cycle for three

river basins. PRECIS-Had, PRECIS-ERA and CRU data show a bell type distribution of

temperature cycle. Generally, in all river basins the characteristics of the mean annual cycle

of temperature in PRECIS RCM simulations are similar to CRU data with highest mean

temperature observed in July and the lowest in January. It is also noted that the highest mean

temperature is apparent in the Astore river basin while the lowest in the Hunza river basin.

This is mainly because of the fact that Hunza river basin is situated at a higher elevation

compared to Astore river basin. Some differences between the PRECIS simulations and CRU

observations are also evident. Differences are present in PRECIS-Had and PRECIS-ERA

simulated temperature as well. However, in some months these differences are minor. The

biases in PRECIS RCM with respect to CRU data are presented in table 4.3. In all river

basins, PRECIS RCM simulations underestimate mean temperature. The magnitude of the

cold bias depends on the driving boundary data and varies from season to season. The cold

bias during winter is somewhat higher in PRECIS-Had compared to PRECIS-ERA while it is

somewhat less in PRECIS-Had compared to PRECIS-ERA during summer. The cold bias in

PRECIS simulations may be because of the deficiencies of the GCM simulations (McGregor,

1997). It is also observed that cold biases in the winter (October to March) are relatively

higher than in the summer (April to September) biases. This is may be because of the fact

that PRECIS RCM simulations give excessive precipitation during OND and JFM seasons

(figure 4.5) resulting into extremely wet soils. Consequently, there is cooling because of high

latent heat flux and low sensible heat flux of wet soils (Bonan, 1998). It is also noted that the

magnitude of cold biases in case of higher altitude Hunza river basins is relatively higher

than in the lower altitude Astore river basin.

4.3.2 Precipitation

For three river basins, the mean annual cycles of precipitation from CRU observations and

PRECIS RCM simulations are shown in figure 4.4. CRU curve shows that the maximum

precipitation is reached in March whereas lowest in September in all river basins. The

PRECIS RCM experimental curves also show similar variability patterns for theses river

basins. Nevertheless, PRECIS RCM simulated precipitation is in general overestimated.


63

There is stronger variability in the mean annual cycle of precipitation when compared with

the mean annual cycle of temperature. PRECIS results also exhibit an overall wet bias, which

is somewhat higher in case of PRECIS-ERA simulation compared to PRECIS-Had

simulation (table 4.4). Moreover, these biases are higher during winter season as compared to

summer season. The wet biases can be best explained by the steep topography of the area,

which can lead to excessive accumulated orographic precipitation (Giorgi, et. al., 1994).

However, some of the biases may be because of the deficiencies in GCMs (McGregor, 1997).

Table 4.3 Biases in mean temperature (˚C) as simulated with PRECIS RCMs relative to CRU reference data for different seasons and river basins (Winter =October- March; Summer =April-September)

Temperature ( ˚C)

PRECIS-Had PRECIS-ERA River Basin

Winter Summer Winter Summer

Astore -5.4 0.4 -3.8 -2.4

Gilgit -9.3 -2.9 -7.2 -4.4

Hunza -9.2 -4.9 -7.1 -6.5

Table 4.4 Biases in precipitation (%) as simulated with PRECIS RCMs relative to CRU reference data for different seasons and river basins (Winter =October- March; Summer =April-September)

Precipitation (%)

PRECIS-Had PRECIS-ERA River Basin

Winter Summer Winter Summer

Astore 135 57 361 133

Gilgit 43 69 103 44

Hunza 167 217 297 210


Figure 4.3

"',""'::; -5;;;:;; -10E~ -15

i::,~

~ -5:;;CLE -10~,f-

E~,::;;;; 0:;;CL

E -5~.f-

15

]CIa iHunza RlVerBas;n

0/' - -::-'>(- ' . -:-'X:.~ "-

,,:,.x., '''''XIo,- -.x' ....

,- -"" ''>/,~ ~

;:../" A,,;,; ""

- ' ' ,x,- "x ' .-- '" ,; ~'; , , , X

/' ',-20

- "

.30

IAN FEB MAY SEP DECOCT NOVJlIN IUL AUGMAR APR

Mdnth

1- - - - PRECIS.fud- CRU. . . X . . . PRECIS-ERA I

15

b) Gilglt River Bas in10

/- -::: :x =-. :--.""",'" ,x' "-

'" " '

,-",~,., ~,'" "'..,' '"

,;.x' .').'-'.T" ~

,'/ ".,' -,x;,/ '.

X '" ~",'" <:,'.

,; ," "x

,"-

,

0

-15

-20

-25

IAN MAR IUN AUG NOV DECSEP OCTAPR MAY IULFEB

Month

1- - - - PRECIS-fud50~ CRU. . . X - - -PRECIS.ERAI

20

15c) As tore R;yerBasll1

10 ~ :--:_--::-.~,"',' " '"

/.,; ,,X' . . 'X" -. .," '"

'" ,x' ',' -,; . '~

,- " x, "-,;,~ ,-(. ' '"..:-~. ~'"

,'/ , "x..-x,- ,x" / ,

/ ./'"

-10

-15

-20

IAN FEB APR' MAY WL SEP DECOCT NOVMAR IUN AUG

Month

1- - - -PREeIS-fud50-eRU . "X' - 'PRECIS-ERAJ

Mean annual cycle of temperature [DC]over (a) Hunza river basin, (b) Gilgitriver basin and (c) Astore river basin as simulated with PRECIS RCMs andfrom CRU data

64


Figure 4.4

=.." ,"'" -E..§. 4.::0

'j 3:9-.e. ,.~ -

cL

1

0

~-"- ',-"

/' x. ,'x '.,

' '."'" '",

x '" ~/ ",

'" "'X. ------'" " ""'><'-"'-:'><""'X' - --

7rI

6 1!

,x..a)Hullza RLverB,HIIL

'-

><

x- - . . . x'

.....--...../

=.

~ 4.S

.Q 3'"

lcIl..

='"~ 8S

:i (,

.§- 4:::

Il..2

0

TAN FEB MAR, , ;

APR MAY TUN TUL AUG SEP OCT NOV DEC

6 r!

5 -;

Month

1- - - -PREClS-H,d-CRD"'X" 'PRECIS-ERA1

..................... .-..-..--..........-.......... -.._---..--..............

b) Gilglt RlverBas 111

x

.- -,, r -.. "

: '" " ,' '" '"

,'", 'x.",>y "" \

.' '" 'x,.' '" \,

x'", ~ "'.I,./"" \ x,'" "

0-1---

.x

TAN MAR

Month

FEB

12 ,""""""

10 .,,,'

"

x

r ;

APR MAY TUN TUL AUG SEP OCT NOV DEC

1- - - - PREClS-.Hod"- CRD. - .X . . . PRECIS-ERA 1

-.......................... --.......................

c)Astore RIVerBas.,.x

;xx

0/< . .

/"' ""' , x.,.. "'~-_./ ',"''''..x' 'X',.' /

I '- - - -, ". .' /'" , ~ 'x --'--

TAN FEB

-----r--i

MAR APR MAY TUN TUL AUG SEP OCT NOV DEC

Month

1- - - -PREcrS.H,dSO-CRD"'X" -PREClS.ERA1

Mean annual cycle of precipitation [mm/day] over (a) Hunza river basin, (b)Gilgit river basin and (c) Astore river basin as simulated with PRECIS RCMsand from CRU data

65


66

4.3.3 Bias correction in PRECIS simulations

Upper Indus Basin is a data sparse region. In some cases, such as Hunza river basin, the daily

meteorological data required for hydrological modelling is not available. Therefore, to

calibrate a hydrological model the necessary information can be drawn from RCMs.

However, biases in RCM simulations may mount to erroneous results of hydrological model

(Akhtar et al., 2008a). Therefore, bias correction in RCM data is deemed necessary to

produce realistic sequence of stream flow (Wood et al., 2004) and the biases in PRECIS

simulations are required to be corrected before applying the temperature and precipitation

data series as input into the hydrological model. An approach as used by Durman et al.

(2001) was applied for the bias correction. In this approach, a monthly factor based on the

ratio of present day simulated value to observed data on a grid box basis is applied to the

modelled climatic variables. Recently, Fowler et al. (2007b) also used this approach to study

the impact of climate change on the water resources in north-west England.

4.4 River Basin Modelling

River basin modelling can be undertaken using various types of hydrological models,

including empirical, conceptual and physically based models. Empirical models are based on

mathematical equations that do not take into account the underlying physical processes and

therefore are not useful for the simulation of various model components. Physically based

models incorporate physical laws based on the conservation of mass, momentum and energy.

The governing equations include lot of parameters and must be solved numerically. The high

amount of parameters may result in different parameter combinations giving equally good

output performances. Moreover, these models generally incorporate too many processes and

too complex formulations at a too detailed scale that is not needed in the context of climate

change. Therefore, the conceptual models seem to be an attractive alternative. These models

are usually able to capture the dominating hydrological processes at the appropriate scale

with accompanying formulations. Therefore, conceptual models can be considered as a nice

compromise between the need for simplicity on the one hand and the need for a firm physical

basis on the other hand. A disadvantage may be that it is generally impossible to derive the

model parameters directly from field measurements and therefore calibration techniques must

be used (Booij, 2002). In this study, HBV model of the Swedish Hydrological and


67

Meteorological Institute (SMHI) has been chosen to study the climate change impact on the

water resources. It has been applied in more than 40 countries and its applications cover

basins in different climatological and geographical regions, ranging in size from less than 1

to more than 40 000 km2 (SMHI, 2005).

4.4.1 Description of HBV model

For river discharge simulation, the hydrological model HBV of the SMHI (Bergström, 1995;

Lindström et al., 1997) is used. The model has been widely used in Europe and other parts of

the world in climate change impact studies (Liden and Harlin, 2000; Bergström et al., 2001;

Menzel and Bürger, 2002; Booij, 2005, Menzel et. al., 2006). Using inputs from RCMs this

model estimated the discharge fairly well for the Suir river in Ireland (Wang et. al. 2006). A

study by Te Linde et al., (2007) compared the performance of two rainfall-runoff models

(HBV and VIC) using different atmospheric forcing data sets and recommended HBV model

for climate change scenarios studies. HBV is a semi-distributed conceptual hydrological

model using sub-basins as the primary hydrological units. It takes into account area-elevation

distribution and basic land use categories (glaciers, forest, open areas and lakes). Sub-basins

are considered in geographically or climatologically heterogeneous basins. The model

consists of six routines, which are a precipitation routine representing rainfall and snow, a

soil moisture routine determining actual evapotranspiration and controlling runoff formation,

a quick runoff routine and a base flow routine which together transform excess water from

the soil moisture routine to local runoff, a transformation function and a routing routine (see

figure 4.5).

The precipitation accounting routine defines actual precipitation (P) as rainfall (RF) or

snowfall (SF) by applying of a threshold value (TT) shown in equation 4.1 and 4.2

respectively.

TTTifPrfcfpcorrRF .. (4.1)

TTTifPsfcfpcorrSF .. (4.2)

where (T) is actual temperature, rfcf is rainfall correction factor, sfcf is snowfall correction

factor and pcorr is precipitation correction factor. In this routine, snowmelt (Sm) is based on

a simple degree-day relation given in equation (4.3). The snow pack is assumed to retain melt


68

water as long as the amount does not exceed a certain fraction of the snow. When

temperature decreases below TT, melt water refreeze (Rmw) according to the equation (4.4).

TTTcfmaxSm (4.3)

).(max. TTTcfcfrRmw (4.4)

where cfmax is a melting factor and cfr a refreezing factor.

Glacier melting (Gm), which occurs only in glacier zones is taken into account by equation

(4.5).

TTTgmeltGm (4.5)

where gmelt is a glacier melting factor.

The soil moisture routine is the main part controlling runoff formation in which direct runoff,

indirect runoff and actual evapotranspiration are generated. Direct runoff occurs if the soil

moisture volume (SM) in the catchment, conceptualised through a soil moisture reservoir

representing the unsaturated soil, exceeds the maximum soil moisture storage denoted by

parameter FC. Otherwise, precipitation infiltrates in the soil moisture reservoir. This

infiltrating precipitation (IN) either replenishes the soil moisture content, seeps through the

soil layer (R) or evapotranspirates. The indirect runoff (R) through the soil layer is

determined by the amount of infiltrating water and the soil moisture content through a power

relationship with parameter BETA, which is shown in equation (4.6).

BETA

FCSMINR

(4.6)

This indicates that indirect runoff increases with increasing soil moisture content. In case of

zero infiltration, indirect runoff also becomes zero. Actual evapotranspiration (Ea) depends

on the measured potential evapotranspiration (Ep), the soil moisture content and parameter

LP which is a limit where above the evapotranspiration reaches its potential value. This is

shown in equations (4.7) and (4.8).

pa EFCLP

SME

FCLPSMif (4.7)

pa EE FCLPSMif (4.8)


69

At the quick runoff routine, three components are distinguished. Components are percolation

to the base flow reservoir, capillary transport to the soil moisture reservoir and quick runoff.

Percolation is denoted by parameter PERC which is a constant percolation rate. This occurs

when water is available in the quick runoff reservoir. Capillary transport is a function of the

maximum soil moisture storage, the soil moisture content and a maximum value for capillary

flow (CFLUX) as shown in equation (4.9). If the yield from the soil moisture routine is

higher than the percolation and the capillary flow, the water becomes available in the quick

runoff reservoir for quick flow which is shown by equation (4.10).

FCSMFCCFLUXC f

(4.9)

ALFAf UZKQ 1

0 (4.10)

where UZ is the storage in the quick runoff reservoir, ALFA a measure for the non-linearity

of the flow in the quick runoff reservoir and Kf a recession coefficient.

The slow flow of the catchment is generated in the base flow routine through equation (4.11).

LZKQ s 1 (4.11)

where LZ is the storage in the base flow reservoir and Ks a recession coefficient.

In the transformation routine, the discharge of each sub-catchment is routed through a

triangular distribution function and in the routing routine the discharges from the sub-

catchments are linked.

Several other parameters such as lapse rate, parameters for temperature, precipitation and

evapotranspiration, forest dependent parameters and snow, lake and glacier parameters can

be used. Furthermore, sub-basins can contain different elevation zones and for each elevation

zone different land use types are considered (the most important are field and forest). Finally,

simplifications such as long-term mean values for evapotranspiration corrected by the actual

temperature can be used instead of measured evapotranspiration.

In order to assess the performance of the model in simulating observed discharge behaviour

an objective function Y is used, which combines the Nash-Sutcliffe efficiency coefficient NS

(Nash and Sutcliffe, 1970) and the relative volume error RE and is defined as


70

RENSY

1

(4.12)

where

Ni

ioo

Ni

ios

QiQ

iQiQNS

1

21

2

-1 (4.13)

Ni

io

Ni

ios

iQ

iQiQRE

1

1100 (4.14)

where i is the time step, N is the total number of time steps, Qs represents simulated

discharge, Qo is observed discharge and oQ is the mean of Qo over the calibration/validation

period. For a favorable model performance, the efficiency NS should be as high as possible

and the RE value should be close to zero.

4.4.2 Model experiments

To study the climate change impacts on the water resources, the influence of meteorological

forcing data on the performance of hydrological model is tested as a first step. Three data sets

used in this test are meteorological stations observations, PRECIS-ERA and PRECIS-Had

bias corrected simulations. Depending on the source of input data (meteorological stations

data, PRECIS-ERA and PRECIS-Had), three HBV models were developed hereafter referred

as HBV-Met, HBV-ERA and HBV-PRECIS, respectively. These models were calibrated and

validated for selected three river basins. The robustness of HBV models was tested by

calibrating the model with one data source and applying the calibrated model to other two

different data sources.


71

Figure 4.5 A schematic diagram of the hydrological model HBV (modified after Lindström et al., 1997), numbers in brackets refer to described equations


72

4.4.3 Calibration and validation of HBV models

During the calibration of HBV-Met, HBV-ERA and HBV-PRECIS models parameters were

estimated with each model for each river basin. The parameters were estimated in two steps.

Firstly, the key parameters were determined. Secondly, a sensitivity analysis was conducted

on the basis of key parameters to obtain optimal parameter set for the HBV-Met, HBV-ERA

and HBV-PRECIS models. The parameters were selected on the basis of physical reasoning,

previous studies (table 4.5) and univariate sensitivity analyses. These parameters are well

documented in the literature (Killingtveit and Saelthun, 1995; Diermanse, 2001; Carr, 2003;

SMHI, 2005; Booij, 2002, 2005) and the values and ranges of some key parameters are

summarized in table 4.5.

For each river basin, a univariate sensitivity analysis was performed to assess the influence of

individual parameter on the output of the model. This was done by varying the value of one

parameter while keeping other parameters constant (default value). Figure 4.6 illustrates the

sensitivity of model parameters. For the three river basins, parameters like gmelt, FC, DTTM,

TT, PERC, and cfmax are found to be most sensitive. There appeared a strong

interdependence among these parameters. In the second step, a multivariate sensitivity

analysis was performed to estimate the most sensitive parameters of the HBV-Met, HBV-

ERA and HBV-PRECIS models for each river basin. Figures 4.7 through 4.9 show the

multivariate sensitivity analysis using these variables (FC, gmelt, TT and DTTM) in the

HBV-Met, HBV-ERA and HBV-PRECIS models for Hunza river basin, respectively. The

values of remaining key parameters (PERC and cfmax) were optimized by univariate

sensitivity analysis. The optimal values of parameters obtained from this sensitivity analysis

are given in table 4.6, while default values of remaining parameters (RFCF, SCFC, PCALT,

ATHORN, TCALT, ALFA, BETA, PCORR and PCALTL) as provided in a study by SMHI

(2005) were used. The value of Hq was derived from measured data following the procedure

given in the same study. Similar analyses were carried out for other two river basins (Gilgit

and Astore). However, figures were not shown in order to avoid redundancy.

Table 4.6 shows the calibrated HBV parameter values for three river basins with three

different input data sets. It shows that the parameter values of HBV-Met, HBV-ERA and

HBV-PRECIS generally fall within the limits described in other studies (Uhlenbrook et


73

al.,1999; Krysanova et al., 1999; SMHI, 2005; Booij, 2002, 2005). However, the parameter

values vary between the three data sources in different river basins. During calibration, it is

noted that for the investigated river basins threshold temperature is the most critical

parameter because PRECIS RCM simulations (figure 4.3 and 4.4) generally show that most

of the precipitation occurs under freezing conditions when the precipitation is in the form of

snow. On the other hand, most of the runoff is generated in summer when temperature is

above the freezing point.

Table 4.7 presents the efficiency (Y), NS, relative volume error and mean observed and

simulated discharge by the three HBV models during calibration and validation periods for

the three river basins. All three HBV models show that the average simulated and observed

discharge is close to each other during the calibration period and consequently the relative

volume error is very small. General testing of conceptual models has shown that NS values

higher than 0.8 are above average for runoff modelling in glaciated catchments (Rango,

1992). Therefore, NS values during calibration are satisfactory for all HBV models and the

highest values are achieved generally by HBV-Met (e.g. 0.67 < NS < 0.87). Figure 4.10

shows the observed and HBV-Met, HBV-ERA and HBV-PRECIS simulated discharge

during calibration period for Hunza river basin. The peak values are in general

underestimated and discharge during low flow period is better simulated by the HBV models.

The double-mass curves constructed from the observed and simulated discharge of Hunza

river basin shows a straight line (figure 4.11). This is an indication of similar behavior of

observed and HBV simulated discharge during calibration. The discharge behavior simulated

by the HBV models for other two river basins has shown similar response (figures are not

given for redundancy). During the calibration period, efficiency (Y) values and visual

inspection of hydrographs demonstrate that the performance of all HBV models is

satisfactory. During validation, the RE values show that in most of the cases all models

underestimate discharge in the three river basins. Overall, only one out of nine combinations

of river basins and HBV models shows a higher efficiency (Y) in the validation period

compared to the calibration mainly due to large relative volume error. The values of the

performance criteria show that during the validation period overall performance of HBV-Met

(e.g. 0.63 < Y < 0.90) is somewhat better compared to the overall performance of HBV

models driven by PRECIS outputs (e.g. 0.42 < Y < 0.81). The efficiency is highest for Hunza


74

river basin compared to the Gilgit and Astore river basins as already observed during

calibration. However, comparison of Y values between different river basins has to be done

carefully because this statistical measure is strongly influenced by runoff variability. This

may explain the relatively low values at Astore river basin, where runoff variability is highest

due to the small size of the river basin. The mean annual cycle of observed discharge and

simulated discharge is shown in figure 4.12. It appears that HBV-PRECIS simulated mean

annual discharge closely follows the pattern of observed discharge i.e. peak discharge in

Hunza river is reached in the month of August and for Gilgit and Astore rivers it is evident in

the month of July. However, peak discharge simulated by the HBV-ERA and HBV-Met

deviates from observed pattern and both model simulated peak discharge in both Gilgit and

Astore rivers in the month of August. It is noteworthy that all HBV models overestimate

discharge at the end of melt season (September-October) and underestimate discharge during

peak flow period (July-August). The extreme events inside the calibrated range are either

overestimated or underestimated and make it difficult to separate the effects of errors in the

input data and model structure (Weerts, 2003). It means that the inherent uncertainty is

enhanced when the models are used outside their calibrated range, which is common practice

in climate scenarios studies. Therefore, to select a model for climate change impact study the

robustness of HBV models is ought to be tested.


75

Table 4.5 Values and range of important parameters found in different studies using HBV model

KHQ FC LP TT DTTM GMELT CFMAX PERC

28-125 0.5-1 -2-2 2.5-7 0-0.1 Carr, (2003)

0.005-0.2 100-1500 <=1 -2-2 -2-2 4 2-4.5 0.01-6 SMHI, (2005)

0.09 1.0 0 -0.5 4 3.0 0.5 SMHId, (2005)

0.01-0.17 100-660 0.2-0.8 0.4-0.8 Booij,(2002, 2005)

0-580 0.8 0 0 4 0.6 Diermanse, (2001)

75-300 0.7-1 -1-2 -1-2 3-6 0.5-1 Killingtveit and Saelthun, (1995)

d = default value

Table 4.6 Parameter values for HBV for three river basins with three different input data sets

HBV-Met HBV-ERA HBV-PRECIS

River basins Parameter

Hunza Gilgit Astore Hunza Gilgit Astore Hunza Gilgit Astore

cfmax 3 3 4.5 3.2 3 3.5 3 3 3.5

DTTM 0 -2.5 -2.5 -1 -1.5 -1.5 -1.5 -2.5 -2.5

FC 1500 700 700 100 700 700 1100 700 700

gmelt 3.5 4 4.5 3.5 3.5 4 4 3.4 4.5

PERC 0.5 0.8 0.8 0.5 0.5 0.5 0.5 0.9 0.5

TT 0 -2 -2.5 -0.3 -1.5 0 0.4 -2 -1.5


76

Table 4.7 Performance of three HBV models during calibration and validation periods in different river basins

Calibration Validation Model Rive Basin

Period Qo Qs NS RE Y Period Qo Qs NS RE Y

Hunza 1981-90 306.5 305.0 0.874 -0.4 0.87 1991-96 280.5 276.6 0.910 -1.4 0.90

Gilgit 1981-90 266.8 265.7 0.825 -0.4 0.82 1991-96 292.6 257.6 0.770 -11.7 0.69

HB

V-M

et

Astore 1981-90 133.4 131.7 0.677 -1.2 0.67 1991-96 171.8 145.7 0.726 -15.2 0.63

Hunza 1981-90 306.5 306.7 0.891 0 0.89 1991-96 280.5 285.3 0.828 1.6 0.81

Gilgit 1981-90 266.8 267.6 0.750 0.2 0.75 1991-96 292.6 261.0 0.759 -10.6 0.69

HB

V-E

RA

Astore 1981-90 133.4 133.3 0.577 0 0.58 1991-96 171.8 132.7 0.515 -22.7 0.42

Hunza 1981-90 306.5 306.5 0.769 0 0.77 1975-80 394.4 294.1 0.696 -25.4 0.56

Gilgit 1981-90 266.8 267.7 0.740 0.3 0.74 1965-70 302.6 275.8 0.758 -8.8 0.70

HB

V-P

REC

IS

Astore 1981-90 133.4 130.2 0.620 -2.3 0.61 1975-80 121.8 127.9 0.622 5.0 0.59


77

Figure 4.6 Sensitivity of HBV model parameters for (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin

FC

LP

rfcf

scfc

TT

DTTM

cfmax

BETA

PERC

ALFA

K4

gmelt

(a) Hunza river basin

FC

LP

rfcf

scfc

TT

DTTM

cfmax

BETA

PERC

ALFA

K4

gmelt

(b) Gilgit river basin

FC

LP

rfcf

scfc

TT

DTTM

cfmax

BETA

PERC

ALFA

K4

gmelt

(c) Astore river basin


78

Figure 4.7 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV-Met (Hunza river basin), with Y as a function of FC,

GMELT, TT and DTTM

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=-1

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=-1

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=3.5DTTM=0

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=0

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=0

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=3.5DTTM=1

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=1

900 1200 1500-1

0

1

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=1


79

Figure 4.8 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV-ERA (Hunza river basin), with Y as a function of FC,

GMELT, TT and DTTM

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=-1

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=-1

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=3.5DTTM=0

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=0

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=0

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=3.5DTTM=1

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=1

100 600 11000.2

-0.3

-0.8

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=1


80

Figure 4.9 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV-PRECIS (Hunza river basin), with Y as a function of FC,

GMELT, TT and DTTM

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=-1

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=-1

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=3.5DTTM=-1.5

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=-1.5

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=-1.5

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=3.5DTTM=-2

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4DTTM=-2

1100 1300 1500-0.1

0.4

0.9

FC

TT

0.85-0.90.8-0.850.75-0.80.7-0.750.65-0.70.6-0.650.55-0.60.5-0.55

GMELT=4.5DTTM=-2


81

Figure 4.10 Observed and simulated discharge (m3/s) of (a) HBV-Met, (b) HBV-ERA and (c) HBV-PRECIS for Hunza river basin during calibration period

0

500

1000

1500

2000

1981 1982 1983 1984 1985 1986 1987 1988 1989

Dis

char

ge (m

3 /s)

Simulated discharge (Qs) Observed discharge (Qo)

a) HBV-Met

0

500

1000

1500

2000

1981 1982 1983 1984 1985 1986 1987 1988 1989

Dis

char

ge (m

3 /s)


b) HBV-ERA

0

500

1000

1500

2000

1981 1982 1983 1984 1985 1986 1987 1988 1989

Dis

char

ge (m

3 /s)


c) HBV-PRECIS


:s

11200~1000.,"

~

.~ 800"'"

]: 600.,"-,.~ 400v,~

{; 200~ a'-''".,:

OJ,~ 1200~7. 10002"...,

"* 800'6

~ 600

] 400'.."'"

~ 200:5Ei'iQ-0:

OJ,

~ 1200S2

E' 1000;7: 800'6

':i: 600~:5E 400'Vi

~ 200...,

:5Ei'i.:;.-0:

a) HBV-Met

a 200 400 600 800 1000 1200

Accum ulated 0 bserved discharge (1J3m3/s)

b) HBV-ERA

aa 400 1000 1200600 800200

Accumulated observed d~scharge (1J3m3/s)

c) HBV-PRECIS

aa 200 400 600 800 1000 1200

Accumulated 0 bserved discharge (1J3m3/s)

Figure 4.11 Double mass-curve analysis relating obser~ed and simulated discharge (m3/s) of(a) HBV-Met, (b) HBV-ERA and (c) HBV-PRECIS for Hunza river basinduring calibration period

82


83

Figure 4.12 Observed and simulated (HBV-Met, HBV-PRECIS and HBV-ERA) mean annual discharge (m3/s) cycle of (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin

0

200

400

600

800

1000

1200


Month

Dis

char

ge (m

3 /s)

HBV-PRECIS Observed HBV-Met HBV-ERA

a) Hunza river basin

0

200

400

600

800

1000


Month

Dis

char

ge (m

3 /s)


b) Gilgit river basin

0

100

200

300

400

500


Month

Dis

char

ge (m

3 /s)


c) Astore river basin


84

4.4.4 Representation of flood peaks

Extreme value analysis based on the Gumbel extreme value distribution is carried out to

estimate the ability of HBV models to simulate flood peaks for three river basins. For this,

the maximum discharge per hydrological year is determined from both measured and

simulated discharges of three river basins. In this analysis, observed discharge data from the

period 1981-1996, HBV-Met discharge data for the same period and HBV-PRECIS

discharge data from the period 1961-1990 are used. Figure 4.13 shows the extreme value

distribution of floods derived from observed discharge data and depicted by HBV-Met and

HBV-PRECIS models. Overall trend of present day simulated annual maximum discharge by

HBV models is an underestimation of flood peaks at all return levels. The highest differences

between observed and modeled extreme discharges are found in the Astore river basin, which

may owe to the small size of the river basin. However, it is difficult to compare the observed

extreme values with simulated extreme values because the extreme discharge return values

are influenced by the period of study. Moreover, observed and HBV-Met simulated extreme

values are based on relatively few extreme flood events, which make the extrapolation to

large return periods highly prone to errors.

4.4.5 Robustness of HBV models

The robustness of HBV models is tested by calibrating the model with one data source and

applying the calibrated model to other two data sources. The Absolute Relative Deviation

(ARD) in the efficiency (Y) is quantified by equation (4.15)

c

ca

YYY

ARD

100 (4.15)

where Yc is the efficiency of the model during calibration and Ya is the efficiency of the

model during the application of a different data source.

The efficiency (Y) and Absolute Relative Deviation (ARD) of the three HBV models using

input data from sources different from the data used in the model calibration are shown in

table 4.8. The values of the efficiencies during the period 1985-86 show that overall

performance of HBV-Met (0.20 < Y < 0.65) is somewhat less compared to the models using

PRECIS RCM data sources (0.31 < Y < 0.86). The ARD values indicate that the errors in


85

HBV-Met (e.g. 18 < ARD < 70) are higher compared to the errors in models using PRECIS

data sources (e.g. 6 < ARD < 46). This can be best explained due to the fact that for each

river basin temperature and precipitation data of only one meteorological station is used as

input to HBV-Met model, which might have lead to extreme behavior and HBV-Met

appeared to be less robust as compared to HBV-ERA and HBV-PRECIS. Moreover, the

values of ARD in case of HBV-PRECIS are somewhat less compared to HBV-ERA values.

This may be attributed to the small precipitation biases in PRECIS-Had compared to

PRECIS-ERA. The application of bias correction only corrects the monthly mean and does

not account for corrections in the variability. Any other sophisticated approach for bias

correction (Leander and Buishand, 2007) may give better results. These results however

indicate that the robustness of HBV models is affected by the input data.

The effect of three different input forcing data series on the simulated discharge of HBV is

analyzed by calculating the uncertainty range. The uncertainty range in a HBV model is the

difference between the maximum and minimum values of the three simulated discharge

series. Figures 4.14-4.16 show the uncertainty range in three HBV models by applying inputs

from three different data sources for three river basins during the 1986 hydrological year.

This is an average hydrological year and is used as an example. The uncertainties are

minimum in the biggest river basin (Hunza river basin) and maximum in the smallest river

basin (Astore river basin). Generally, the uncertainties are higher during the start (Aprial-

May) and end (August-September) of the melt season. During the peak flows, uncertainties

are somewhat less in HBV-PRECIS compared to the HBV-Met and HBV-ERA. These

results indicate that forcing data largely influence the performance values and HBV-PRECIS

performed better compared to HBV-Met and HBV-ERA in terms of robustness and

uncertainty. The uncertainties in the three HBV models show that three forcing data series

have a large influence on the simulated discharge. The uncertainty range varies among the

three river basins. The uncertainties are somewhat less in the Hunza river basin compared to

the Gilgit and Astore river basins. This may be due to the fact that the Hunza river basin is

heavily glaciated (34 %) and temperature play a major role in the summer discharge, whereas

the discharge of less glaciated Gilgit (7 %) and Astore (16 %) river basins depends on the

preceding winter precipitation (Archer, 2003). Since in the three different forcing data sets

the temperature series are stable compared to the precipitation series, the bias correction


86

technique applied has a larger impact on the precipitation series compared to the temperature

series. Consequently, there are less uncertainties in the simulated discharge in the Hunza

river basin compared to Gilgit and Astore river basins.

Table 4.8 Efficiency Y of three HBV models using data sources different from the

calibration sources during the hydrological years 1985 and 1986 in different river basins. The values of absolute relative deviations (ARD) are given in parentheses. The italic values indicate efficiency Y during calibration

Data Source Applied River Basin Model

Met Observations

PRECIS-ERA PRECIS-Had

HBV-Met 0.87 0.65(25) 0.53(39)

HBV-ERA 0.49(45) 0.89 0.73(18) Hunza

HBV-Had 0.56(27) 0.86(12) 0.77

HBV-Met 0.82 0.55(33) 0.62(24)

HBV-ERA 0.57(24) 0.75 0.63(16) Gilgit

HBV-Had 0.67(9) 0.64(14) 0.74

HBV-Met 0.67 0.20(70) 0.55(18)

HBV-ERA 0.31(46) 0.58 0.37(36) Astore

HBV-Had 0.57(6) 0.35(43) 0.61


87

Figure 4.13 Observed, HBV-Met simulated and HBV-PRECIS simulated annual maximum discharge as a function of return period for three river basins in the present day climate

0

500

1,000

1,500

2,000

2,500

3,000

-2 -1 0 1 2 3 4

Reduced Gumbel variate

Dis

char

ge (m

3 /s)

HBV-Met

HBV-PRECIS

Observed

a) Hunza River Basin

0

500

1,000

1,500

2,000

2,500

3,000

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)

HBV-Met

HBV-PRECIS

Observed

b) Gilgit River Basin

0

500

1000

1500

2000

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)

HBV-Met

HBV-PRECIS

Observed

c) Astore River Basin

1 2 5 10 20 50

Return period (years)

1 2 5 10 20 50


1 2 5 10 20 50



88

Figure 4.14 Observed discharge (green line) and uncertainties in discharge (red shade)

simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Hunza river basin during the 1986 hydrological year

0

500

1000

1500

2000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

1500

2000c) HBV-PRECIS

0

500

1000

1500

2000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

1500

2000b) HBV-ERA

0

500

1000

1500

2000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

1500

2000a) HBV-Met


89

Figure 4.15 Observed discharge (green line) and uncertainties in discharge (red shade)

simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Gilgit river basin during the 1986 hydrological year

0

500

1000

1500

2000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

1500

2000c) HBV-PRECIS

0

500

1000

1500

2000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

1500

2000b) HBV-ERA

0

500

1000

1500

2000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

1500

2000a) HBV-Met


90

Figure 4.16 Observed discharge (green line) and uncertainties in discharge (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Astore river basin during the 1986 hydrological year

0

500

1000

1 31 61 91 121 151 181 211 241 271 301 331 3610

500

1000a) HBV-Met

0

500

1000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

b) HBV-ERA

0

500

1000

1 31 61 91 121 151 181 211 241 271 301 331 361

Day

0

500

1000

c) HBV-PRECIS


91

4.5 Summary

It appears that hydro meteorological variables pertaining to the Upper Indus Basin are

strongly influenced by altitude. Difficult topography makes most of the region inaccessible

for routine meteorological and climatological observations and data are scarce. Most of the

observed hydro meteorological parameters, required for hydrological modelling, are not

available. The observed meteorological records show a reduction of mean temperature during

heavy rainfall events, which leads to a sharp decrease in discharge of highly glaciated Hunza

river basin. However, the discharge of Gilgit and Astore rivers appears to be less sensitive to

this condition. This is mainly because of the fact that discharge of these river basins is highly

correlated with preceding winter precipitation (Fowler and Archer 2005). In contrast,

discharge in Hunza river basin is uncorrelated with winter precipitation but highly correlated

with summer mean temperature (Archer 2003).

The temporal patterns of the simulated mean annual cycle of temperature and precipitation

are similar to CRU data in all river basins. However, some quantitative differences between

PRECIS simulations and CRU data also exist. Generally, PRECIS simulations underestimate

temperature and overestimate precipitation with respect to CRU data, which is a common

phenomena observed in mountain areas (Giogori, 2002; Kumar et al., 2006; Solman et al.,

2008). The biases are highly influenced by the driving forcing data. Overall, in three river

basins the magnitude of temperature biases is somewhat higher in PRECIS-Had compared to

PRECIS-ERA simulation whereas the magnitude of precipitation biases is somewhat less in

PRECIS-Had compared to PRECIS-ERA simulation. Application of this data in hydrological

impact studies without bias correction may lead to wrong interpretation because during the

calibration, biases in RCM simulations might adjust parameters of hydrological model in

such a way that lead to false results (Akhtar et al., 2008a). Therefore, bias correction of RCM

data is deemed necessary to produce realistic sequence of stream flow (Wood et al., 2004).

Hence, the biases observed in PRECIS simulations are corrected before applying the

temperature and precipitation data series as input into the hydrological model by following

the approach of Durman et al., (2001).

The calibration and validation results of the HBV hydrological model driven by observed

data and PRECIS RCM present day simulations show that the HBV model can reproduce the


92

discharge reasonably well. In terms of performance criteria, HBV calibrated with observed

station data simulates discharge behaviour somewhat better than HBV calibrated with

PRECIS RCM data. During the validation period, overall performance of HBV-Met is also

somewhat better compared to the overall performance of HBV models driven by PRECIS

outputs. All three HBV models overestimate discharge at the end of the melt season and

underestimate discharge during the peak flow period. Using the input data series from

sources different from the data used in the model calibration shows that HBV models

calibrated with PRECIS output generally have higher efficiency (Y) and lower absolute

relative deviation (ARD) values compared to the corresponding values of HBV-Met. This

indicates that HBV-Had and HBV-ERA are more robust compared to HBV-Met model. The

patterns of uncertainties are similar in the three HBV models. The magnitude of uncertainties

is higher in the river basins where discharge is dependent on the preceding winter

precipitation (i.e. Gilgit and Astore river basins) compared to the river basin where discharge

is driven by energy inputs (i.e. Hunza river basin). This is may be because of the fact that the

bias correction technique applied here has a larger impact on the precipitation series

compared to the temperature series that resulted in smaller uncertainty in the simulated

discharge of the Hunza river basin. In terms of both robustness and uncertainty ranges, the

HBV models calibrated with PRECIS output performed better compared to HBV-Met.

Therefore, it is recommended that in data sparse regions like the HKH region data from

regional climate models may be used as input in hydrological models for climate scenarios

studies.

Chapter 5 Climate Change Impact on Water Resources

93

CHAPTER 5

CLIMATE CHANGE IMPACT ON WATER RESOURCES

5.1 Background

Climate change may affect, one way or other, the human being and ecosystem. It may

influence agricultural areas because of shifts in growing seasons and/or increase in water

demand (Doll, 2002). It may also has negative effect on the river flow due to intensification

of water cycle and higher frequency of flood events (Milly et al., 2002; Huntigton, 2006;

Kundzewicz et al., 2007). In developing countries, like Pakistan, climate change may

transmit an additional stress on socioeconomic system, which is already under tremendous

pressure due to several factors, including fast population growth rate, rapid urbanization and

fierce economic competition.

The tremendous importance of water for both society and nature emphasize the need of

evaluating the climate change impact on the future water resources of Pakistan. This is

accessed by applying the climate change scenario to the previously calibrated/validated

hydrological model. Future water resources and flood peaks are estimated under SRES B2

scenario at different stages of deglaciation. Different methods as direct and delta approach

are used to simulate discharge for the future climate.

5.2 Change of Temperature and Precipitation in the Selected River Basins

For three river basins, the mean annual cycles of temperature and precipitation for the present

and future climate simulated with PRECIS are presented in Figs. 5.1 and 5.2, respectively. A

general increase in temperature and precipitation during the period 2071-2100 is evident. The

northern river basins (i.e. Hunza and Gilgit river basins) experience more warming relative to

the southern river basin (i.e. Astore river basin). For the southern river basin, larger increases

in precipitation are expected as compared to the northern river basins. Table 5.1 presents the

seasonal changes in temperature and precipitation in the three river basins with climate

change. The annual mean temperature rise by the end of the century ranges from 0.83 to 3.05

˚C. The warming is more pronounced in the Hunza (1.80˚C) and Gilgit (3.09˚C) river basins

when compared with the Astore (0.83˚C) river basin. In the Astore river basin, the weak


94

temperature change signals may be because of the fact that in this basin future PRECIS RCM

simulations give excessive precipitation changes (figure 5.2), which tend to result in

excessively wet soils causing high latent heat and low sensible heat fluxes. Consequently,

surface cooling (Bonan, 1998).

PRECIS estimates a rise in annual mean precipitation (6 to 23%) by the end of the 21st

century. The increase in precipitation is observed in all seasons. The precipitation changes in

the Hunza (6%) and Gilgit (9%) river basins are somewhat similar, while precipitation

changes in the Astore (23%) river basin are comparatively large. The excessive increase in

precipitation in the Astore river basin may be because of an increase in monsoon activities

(southern Astore river basin experiences more monsoon influence compared to northern

Hunza and Gilgit river basins). Generally, the magnitude of predicted temperature and

precipitation change in SRES B2 scenario is somewhat less compared to SRES A2 scenario

as reported by Akhtar et al., (2008a). However, the spread of change is different in SRES B2

as compared to SRES A2 scenario. This may be due to the fact that the domain and

resolution of experiments used in this study are different from that of Akhtar et al., (2008a).

The increase in temperature and precipitation is overall consistent with the projected increase

in temperature and precipitation reflected in SRES B2 scenario of neighboring areas such as

southwest China and northwest India (Yinlong et al., 2006; Kumar et al., 2006).

Table 5.1 Seasonal changes of mean temperature and precipitation under PRECIS simulated SRES B2 scenario for the period 2071-2100 over three river basins relative to the period 1961-1990 (Summer = April-September; Winter=October-March)

Temperature Change (˚C) Precipitation Change (%) River Basins

Annual Winter Summer Annual Winter Summer

Hunza 1.80 2.01 1.59 6 7 3

Gilgit 3.09 3.05 3.14 9 12 4

Astore 0.83 0.92 0.75 23 28 17


95

Figure 5.1 Mean annual cycle of temperature [˚C] over river basins (a) Hunza, (b) Gilgit

and (c) Astore as simulated with PRECIS for present (1961-90) and future (2071-2100) climate under SRES B2 scenario

-30-25-20-15-10-505

10

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month

Mea

n te

mpe

ratu

re (˚

C)



-30

-20

-10

0

10

20


Month

Mea

n te

mpe

ratu

re (˚

C)



-20-15-10-505

101520


Month

Mea

n te

mpe

ratu

re (˚

C)




96

Figure 5.2 Mean annual cycle of precipitation [mm/day] over river basins (a) Hunza, (b)

Gilgit and (c) Astore as simulated with PRECIS for present (1961-90) and future (2071-2100) climate under SRES B2 scenario

0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)



0

2

4

6

8

10


Month

Pre

cipi

tatio

n (m

m/d

ay)



0

2

4

6

8

10

12


Month

Pre

cipi

tatio

n (m

m/d

ay)




97

5.3 Climate Change Signals Transfer from PRECIS RCM to HBV

The interpretation of climate change in terms of hydrological change is not a straightforward

process as meteorological variables derived from climate models are often subject to

systematic errors. Therefore, a reliable interface mechanism is required to transfer results

from RCM to hydrological impact model. There are two commonly practiced approaches,

including delta change approach and scaling approach (Graham et al., 2007b; Lenderink et

al., 2007; Fowler et al., 2007a,b). In another approach, Akhtar et al., (2008a) calibrated a

hydrological model with RCM data without applying any bias correction. In this method,

there is no need to scale the future climate data series. This approach is not yet extensively

tested and there is a risk that potential biases in RCM simulations may lead to over

parameterization. Hence, in this study only delta change and scaling approaches are applied

to assess the climate change impact on future water resources.

5.3.1 Delta change approach

The delta change approach has already been used in many climate change impact studies

(Arnell, 1998; Gellens and Roulin, 1998; Middelkoop et al., 2001). In this approach, the

observed climate data series are adapted with estimated monthly climate changes from the

PRECIS RCM. The observational database used for the delta change approach covers the

period 1981-1996. This is the same period to which HBV -Met has been calibrated and

validated. The future daily temperature ( dailyfT , ) and daily precipitation ( dailyfP , ) time series

are constructed by using equations 5.1 and 5.2, respectively.

monthlypmonthlyfdailyodailyf TTTT ,,,, (5.1)

monthlyp

monthlyfdailyodailyf P

PPP

,

,,, (5.2)

Where dailyoT , is the observed daily temperature, dailyoP , is the observed daily precipitation,

monthlyfT , is the mean monthly PRECIS simulated future temperature, monthlypT , is the mean

monthly PRECIS simulated present temperature, monthlyfP , is the mean monthly PRECIS


98

simulated future precipitation and monthlypP , is the mean monthly PRECIS simulated present

precipitation.

5.3.2 Scaling approach

The delta change approach does not include changes in variability between RCM present and

future scenario simulations. Therefore, to use information derived from climate models

optimally while producing reasonable hydrological simulations is to use a scaling approach.

Scaling implies an adjustment of specific variables to reduce systematic biases. The scaling

factors derived from the present day simulation of a particular climate model are applied to

adjust future scenarios simulated from the same RCM, with the aim of altering RCM results

as little as possible (Graham et al., 2007b; Fowler et al., 2007b; Lenderink et al., 2007). The

future precipitation and temperature data series are corrected with the same bias correction

factor that is used to correct the present day temperature and precipitation data.

5.4 Assessment of Water Resources under Climate Change

We have estimated the future water resources from previously calibrated and validated HBV-

Met and HBV-PRECIS models using both the delta change and scaling approaches. The

effect of climate change on river discharge is simulated for the current glacier extent (100 %

glacier scenario) and for two stages of deglacierisation, i.e. after an areal reduction by 50%

(50 % glacier scenario) and after complete melting (0 % glacier scenario).

5.4.1 Simulation of annual discharge cycle

Figure 5.3 shows the mean annual discharge cycle simulated by HBV-Met for the present

and future climate for three stages of glacier coverage: 100 % glaciers, 50 % glaciers and 0 %

glaciers. The amplitude of the annual discharge cycle is increased in a changed climate under

the 100 % glacier scenario. Snow melting starts one month earlier and discharge rises

towards its peak in summer (August). However, this case has to be regarded as a hypothetical

one because future 100 % glacier extent is not realistic with climate change. If the glacierised

area is reduced by 50%, the discharge is decreased during peak flow season. However, in

Astore river basin snowmelt still begins one month earlier and discharge is increased during


99

the month of March. The monthly runoff for the 0 % glacier scenario is reduced drastically in

all three river basins.

The present and future HBV-PRECIS simulated discharge assuming a glacier coverage of

100 %, 50 % and 0 % are given in figure 5.4. The simulated seasonal discharge pattern

appears to be generally similar to the pattern derived through HBV-Met. The amplitude of

the seasonal discharge cycle is increased in a changed climate under the 100 % glacier

scenario as well. Snowmelt starts one month earlier. There is an increase in river discharge

throughout the year in all river basins. The highest peak is observed in August. If the

glacierised area is reduced by 50%, discharge is decreased in all rivers. The reduction in

discharge mainly predominates during the months of July and August. After complete

reduction of the glaciers, there is a considerable decrease in discharge.

Table 5.2 presents the mean relative changes in future discharge (2071-2100) in a changed

climate relative to the present discharge (1961-1990) for the three glaciations stages in three

river basins. There is a big discrepancy between the results of changes in discharge simulated

by HBV-Met and HBV-PRECIS. Under the 100 % glacier scenario, both models predict an

increase in water resources. Whereas under 50% glacier scenario, the discharge is predicted

to be decreased. The magnitude of predicted increase is higher in HBV-PRECIS compared to

HBV-Met estimates whereas the magnitude of decrease is higher in HBV-Met compared to

HBV-PRECIS. Without glaciers, HBV-Met and HBV-PRECIS predict a drastic decrease in

the discharge up to 96 % and 93 %, respectively. There is neither forest nor any major lake

present in the three river basins and glaciers and fields (area without forest) are considered as

the only two land use classes in the hydrological model framework. Therefore, the effect of

complete melting of glaciers on the hydrological cycle will depend on the degree of

glaciation in the river basins and response of the river basins to climate change. For instance,

looking at the patterns of climate change in the three river basins the highly glaciated Hunza

river is expected to react more severely compared to the least glaciated Gilgit river basin.

However, HBV-Met shows that more drastic changes are expected in the Gilgit river basin

compared to the Hunza river basin. This may be because of the inaccurate transfer of climate

change signals through the delta change approach.


100

Figure 5.3 Annual discharge cycle simulated by HBV-Met for the present climate and future SRES B2 climate for three stages of glaciation for three river basins

0

500

1000

1500


Month

Dis

char

ge (m

3 /s)

Future Discharge-100% glaciation Future Discharge-50% glaciation

Future Discharge-0% glaciation Present simulated discharge


0

500

1000

1500


Month

Dis

char

ge (m

3 /s)




0

100

200

300

400

500


Month

Dis

char

ge (m

3 /s)





101

Figure 5.4 Annual discharge cycle simulated by HBV-PRECIS for the present climate and

future SRES B2 climate for three stages of glaciation for three river basins

0

500

1000

1500

2000


Month

Dis

char

ge (m

3 /s)




0

500

1000

1500


Month

Dis

char

ge (m

3 /s)




0

100

200

300

400

500


Month

Dis

char

ge (m

3 /s)





102

Table 5.2 Mean relative change in future discharge (2071-2100) in a changed SRES B2 climate relative to the present discharge (1961-1990) for three glaciations stages and for three river basins

Mean change in discharge (%) Model River Basin

Under present day glaciations

Under 50 % reduction in glaciers

Without glaciers

Hunza 26 -27 -81

Gilgit 44 -30 -96

H

BV

-Met

Astore 17 -27 -72

Hunza 65 -13 -91 Gilgit 40 -20 -78

HB

V-P

REC

IS

Astore 46 -24 -93

5.4.2 Future discharge peaks

Extreme value analysis based on the Gumbel extreme value distribution is carried out to

estimate the impact of climate change on floods under SRES B2 scenario for three river

basins with two HBV models and for three glaciation stages. For this, the maximum

discharge per hydrological year is determined from both HBV-Met and HBV-PRECIS

simulated discharge series under SRES B2 scenario in three river basins.

The flood frequency results under climate change for three glacier stages estimated through

HBV-PRECIS and HBV-Met are presented in figure 5.5 and 5.6, respectively. In all river

basins, both HBV models show an increase in flood magnitude for all return periods in a

changed climate under the 100 % glacier scenario. These results are in agreement with the

study of Milly et al. (2002) who found an overall increase in flood peaks during the twentieth

century and this trend is expected to continue in the future. The magnitude of flood frequency

under climate change in the 50 % and 0 % glacier coverage stages is decreased in all three

river basins. The change in peak discharge in the HBV-PRECIS at 20-year return level in the

100 % glacier stage is 20 %, 12 % and 6 % in the Hunza, Gilgit and Astore river basins,

respectively. For the 50% and 0% glacier scenarios the flood peaks at 20 year return level


103

decreased in the Hunza (40 % and 91 %, respectively) Gilgit (31 % and 66 %, respectively)

and Astore (36 % and 81 %, respectively) river basins. The change in peak discharge in the

HBV-Met at 20-year return level in the 100 % glacier stage is 18 %, 32 % and 9 % in the

Hunza, Gilgit and Astore river basins, respectively. For the 50% and 0% glacier scenarios the

flood peaks at 20 year return level decreased in the Hunza ( 28 % and 70 %, respectively)

Gilgit ( 35 % and 89 %, respectively) and Astore ( 18 % and 31 %, respectively) river basins.

The characteristics of future annual maximum discharge values under climate change are

given in table 5.4. There are huge outliers in HBV-Met simulated future annual maximum

discharge values in the Hunza river basin. The outliers in HBV-Met are explained by the fact

that in each river basin only one meteorological station is used for temperature and

precipitation input into HBV-Met. Observed precipitation is considered as areally averaged

precipitation but actually point precipitation. Unfortunately, sufficient precipitation stations

were not available to assess the areally averaged basin scale precipitation in a right way.

Consequently, observed precipitation shows too much variability and extreme behavior.

Parameters are estimated under variable and extreme conditions. For example, in Hunza river

basin at Skardu meteorological station there are three heavy rainfall spells in the month of

October 1987 (average rainfall is 37.0 mm in October, 1987 while the climate normal for

October is 6.4 mm). When we use climate change scenarios derived from PRECIS (in

October there is an increase in precipitation of 57 %), HBV-Met gives extremely high peaks

in October 1987 (an increase in mean discharge of 289 % in October 1987). Therefore, the

quality of input data used in HBV-Met seems to be too poor to simulate extreme discharge

behavior.

The modeled changes in flood frequency under climate change are just estimations that are

based on simulations using input data from only one RCM run using one emission scenario

and single GCM for the boundary data. Other GCMs could result in quite different flood

frequency predictions. For instance, the flood frequency changes in this study under SRES

B2 scenario are different from the estimates given by Akhtar et al., (2008a) under SRES A2

scenario. It also appears that assessment method significantly affect the future predictions.

Despite all uncertainties, the behavior of peak discharges predicted by the two HBV models

supports the direct use of RCM output as input to hydrological models in this area.


104

Figure 5.5 HBV-PRECIS simulated annual maximum discharge as a function of return

period for current and changed SRES B2 climate for three glacier stages for three river basins

0

500

1000

1500

2000

2500

3000

3500

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)

Future-100% glaciation Future-50% glaciation

Future-0% glaciation Present simulateda) Hunza River Basin

1 2 5 10 20 50

0

500

1000

1500

2000

2500

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)

Future-100% glaciation Future-50% glaciat ion

Future-0% glaciation Present simulated


0

500

1000

1500

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)

Future-100% glaciat ion Future-50% glaciat ion

Future-0% glaciation Present simulated


1 2 5 10 20 50


1 2 5 10 20 50




105

Figure 5.6 HBV-Met simulated annual maximum discharge as a function of return period

for current and changed SRES B2 climate for three glacier stages for three river basins

0

500

1000

1500

2000

2500

3000

3500

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)


Future-0% glaciat ion Present simulateda) Hunza River Basin

1 2 5 10 20 50

0

500

1000

1500

2000

2500

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)


Future-0% glaciat ion Present simulatedb) Gilgit River Basin

0

500

1000

1500

-2 -1 0 1 2 3 4


Dis

char

ge (m

3 /s)


Future-0% glaciation Present simulatedc) Astore River Basin

1 2 5 10 20 50


1 2 5 10 20 50




106

Table 5.3 Characteristics of future annual maximum discharge simulated by two HBV models in a changed SRES B2 climate for the three glaciations stages and for three river basins. The values in parentheses are future annual maximum discharge with outliers

Means (m3/s) Standard Deviation (m3/s)

Glaciation Scenario Model River Basin

100 % 50 % 0 % 100 % 50 % 0 %

Hunza 1773.9 890.5 83.0 108.4 57.5 33.7

Gilgit 1083.5 633.8 246.3 197.1 182.6 126.1

HB

V-P

REC

IS

Astore 521.6 278.6 43.0 20.3 30.4 28.5

Hunza 1573.6 (1868.7)

866.5 (1182.3)

252.4 (500.4)

256.6 (1169.5)

298.6 (1256.7)

346.5 (1016.8)

Gilgit 1282.7 605.1 34.6 144.3 83.1 40.3

HB

V-M

et

Astore 533.1 357.6 215.7 28.4 49.7 73.3

5.5 Summary There is a general increase in temperature and precipitation towards the end of the 21st

century in the three river basins. Climate change under SRES B2 scenario is showing

reduced magnitude of temperature and precipitation change relative to changes in SRES A2

scenario (Akhtar et al., 2008a). Among the three river basins, the amplitude of temperature

change is highest in Gilgit river basin whereas it is lowest in Astore rivers basin. The annual

mean temperature rise in the three rivers basins by the end of the century ranges between

0.83-3.09 ˚C. In these river basins, the precipitation change ranges from 6 to 23%.

In a changed climate, HBV does not calculate the new glacier area size automatically. To

bridge this deficiency, we have used three glacier coverage scenarios as applied by Hagg et

al. (2007) while modelling the hydrological response to climate change in glacierized Central

Asian catchments. However, future glacier extent may be predicted separately by using a

simple hypsographic modelling approach (Paul et al. 2007). The use of such a predicted


107

future glacier extent in HBV would give a more realistic hydrological change. To quantify

the future water resources, the delta change approach is used for HBV-Met and direct use of

PRECIS RCM data is done for HBV-PRECIS. There are differences in the results of both

approaches. In a changed climate, the discharge will generally increase in both HBV-

PRECIS and HBV-Met in the 100 % glacier coverage stage up to 65% and 44%,

respectively. At the 50 % glacier coverage stage, the discharge is expected to reduce up to 24

% and 30 % both in HBV-PRECIS and HBV-Met, respectively. For the 0 % glacier coverage

a drastic decrease in water resources is predicted by HBV-Met (up to 96 %) and HBV-

PRECIS (up to 93%).

There are huge outliers in annual maximum discharge simulated with HBV-Met. This shows

that the prediction of hydrological conditions through the delta change approach is not ideal

in the Hindukush-Karakorum-Himalaya region. HBV-PRECIS provides results on

hydrological changes that are more consistent with RCM changes. This shows that the

climate change signals in HBV-PRECIS are transmitted more realistically than in HBV-Met.

Therefore, the direct use of RCM outputs in a hydrological model may be an alternative in

areas where the quality of observed data is poor. The direct use of RCM outputs (HBV-

PRECIS model) has shown that the magnitude of annual maximum flood peaks is likely to

increase in the future. Hence, overall results are indicative of a higher risk of flood problems

under climate change. The modeled changes in future discharge and changes in flood

frequency under climate change are not conclusive because more research is needed to

evaluate the uncertainties in this approach. Moreover, this technique needs to be tested with

other RCMs and preferably to river basins in other parts of the world as well.

Chapter 6 Conclusions and Recommendations for Future Work

108

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK

6.1 Conclusions

Following conclusions are drawn from this study.

The spatial patterns of mean temperature and precipitation simulated through PRECIS

RCM agree reasonably well with observed data (CRU data) despite of inherent biases

in modeled data. For example, model results show an overestimation of precipitation

and an underestimation of mean temperature over the high relief (mountainous)

regions.

PRECIS RCM simulations generally overestimate the interannual variability in

temperature for all seasons except for the July-August-September (JAS) season. The

standard deviations in the PRECIS-Had and HadAM3P simulated temperature are

close to each other, a sign that the nested model is inheriting a good proportion of its

temperature interannual variability from HadAM3P. The interannual variability of

precipitation is overall higher during the winter compared to the summer. This owes

to the fact of dividing the standard deviation with very low precipitation value.

Compared to the observational data (CRU data), PRECIS-Had and HadAM3P

simulations show an underestimation of interannual variability in the high

mountainous regions and overestimation in relatively flat regions. The PRECIS-Had

and HadAM3P coefficients of variations are generally in good agreement but

PRECIS-Had some time shows slightly overestimation. This fact leads to the

conclusion that PRECIS-Had data are highly sensitive to the quality of boundary data.

Future scenario, SRES B2, shows an increase in temperature and precipitation by the

end of 21st century relative to the present day climate. The spatial patterns of mean

temperature changes show that warming signals are weak over the mountainous

region. Taking into account only land points, the average annual predicted increase in

temperature and precipitation is 3.1 ˚C and 8 %, respectively. The winters are

envisaged to be warmer than summer. The annual mean temperature rise in the


109

selected rivers basins (Hunza, Gilgit and Astore river basins) is expected to be in

range of 0.83 to 3.09 ˚C whereas precipitation change might vary between 6 to 23%.

In Hunza, Gilgit and Astore river basins, PRECIS RCM simulations underestimate

temperature and overestimate precipitation with respect to CRU data. Overall, the

magnitude of temperature biases is somewhat higher in PRECIS-Had compared to

PRECIS-ERA simulation whereas the magnitude of precipitation biases is somewhat

less in PRECIS-Had compared to PRECIS-ERA simulation. The biases in PRECIS

RCM simulations may lead to some serious discrepancies in hydrological impact

studies. For example, biased data may influence parameters of hydrological model in

a way leading to erroneous results. Therefore, bias correction to PRECIS RCM data

in terms of temperature and precipitation is necessary before applying that data as

input in the hydrological model.

The calibration and validation results of the HBV hydrological model driven by

observed data and PRECIS RCM present day simulated data show that the HBV

model can reproduce the discharge reasonably well. In terms of performance criteria,

HBV model calibrated with observed data simulates discharge behaviour somewhat

better than HBV calibrated with PRECIS RCM data. During the validation period,

overall performance of HBV-Met is also fairly good compared to the overall

performance of HBV models driven by PRECIS outputs. However, all three HBV

models overestimate discharge at the end of the melt season and underestimate

discharge during the peak flow period.

Using the input data series from sources different from the data used in the model

calibration shows that HBV models calibrated with PRECIS output generally have

higher efficiency (Y) and lower absolute relative deviation (ARD) values compared to

the corresponding values of HBV-Met. This indicates that HBV-Had and HBV-ERA

are more robust compared to HBV-Met model.

The patterns of uncertainties are similar in the three HBV models. The magnitude of

uncertainties is higher in the river basins where discharge is dependent on the

preceding winter precipitation (Gilgit and Astore river basins) compared to the river

basin where discharge is driven by energy inputs (Hunza river basin). This owes to


110

the fact that the bias correction technique applied has a notable impact on the

precipitation data compared to the temperature data that resulted in smaller

uncertainty in the simulated discharge of the Hunza river basin.

In terms of both robustness and uncertainty ranges, the HBV models calibrated with

PRECIS output performed better compared to HBV model calibrated with observed

data. Therefore, it is recommended that in data sparse regions like the HKH region,

data from regional climate models may be used as input in hydrological models for

climate scenarios studies.

To quantify the future water resources, the delta change approach is used for HBV-

Met and direct use of PRECIS RCM data is done for HBV-PRECIS. There are

differences in the results of both approaches. In a changed climate, the discharge will

generally increase in both HBV-PRECIS and HBV-Met in the 100 % glacier

coverage stage up to 65% and 44%, respectively. At the 50 % glacier coverage stage

the discharge is expected to reduce up to 24 % and 30 % both in HBV-PRECIS and

HBV-Met, respectively. For the 0 % glacier coverage a drastic decrease in water

resources is predicted by HBV-Met (up to 96 %) and HBV-PRECIS (up to 93%). At

100 % glacier coverage the magnitude of flood peaks is likely to increase in the future

which is an indication of higher risk of flood problems under climate change.

There are huge outliers in annual maximum discharge simulated with HBV-Met. This

shows that the prediction of hydrological conditions through the delta change

approach is not ideal in the HKH region. HBV-PRECIS provides results on

hydrological changes that are more consistent with RCM changes. This shows that

the climate change signals in HBV-PRECIS are transmitted more realistically than in

HBV-Met. Therefore, the direct use of RCM outputs in a hydrological model may be

an alternative in areas where the quality of observed data is poor.


111

6.2 Recommendations for Future Research Work

Recommendations for future work are given as below:

The modeled changes in future discharge and changes in flood frequency under

climate change are not conclusive because more research is needed to evaluate the

uncertainties in the assessment methods. Moreover, the modelling technique is

required to be tested with other RCMs and preferably to river basins in other parts of

the world as well. Uncertainties in climate change projections may result from

different sources including future emissions, model parameterisation and natural

climate variability. Further work must concentrate to examine the uncertainties

associated with the whole climate system modelling and in hydrological impact

modelling.

In a changed climate, HBV does not calculate the new glacier area size automatically.

To bridge this deficiency, we have used three hypothetical glacier coverage scenarios.

However, future glacier extent may be predicted separately by using a simple

hypsographic modelling approach. The use of such a predicted future glacier extent in

HBV would give a more realistic hydrological change. Therefore, further work on the

future glacier coverage modelling is needed.

The use of RCM data in hydrological model is tested only for snow and glacial melt

river basins. Therefore, it is recommended that the modelling techniques applied here

should also be tested for some rain fed river basins. In this study, only one GCM and

one emission scenario have been applied, further work can be extended by using other

GCMs and other emission scenarios.

References

112

REFERENCES

Akhtar, M., Ahmad, N, Chaudhry, M. N, and Babur, K. 2005a. Spatial and temporal variations in precipitation and temperature in the Upper Indus Basin, Pakistan, Proceedings of International Conference Environmentally Sustainable Development (ESDev-2005), 25-27 June, 2005, COMSATS, Institute of Information Technology, Abbotabad- Pakistan, 639-651.

Akhtar, M., Ahmad, N, Chaudhry, M. N, and Hussain, S.P. 2005b. Climate change in the Upper Indus Basin of Pakistan: A Case Study, proceedings of National Workshop on “ Global change prospective in Pakistan–challenges, impacts, opportunities and prospects” 28-30 April, 2005, Pakistan Academy of Science (PAS), Islamabad-Pakistan, edited by Dr Amir Muhammad and Dr. Sajidin Hussain, 14-28.

Akhtar, M., Ahmad, N., and Booij, M.J., 2008a. The impact of climate change on the water resources of Hindukush-Karakorum-Himalaya region under different glacier coverage scenarios. Journal of Hydrology. 355, 148-163.

Akhtar, M., Ahmad, N. and Booij, M.J.2008b. Use of regional climate model simulations as input for hydrological models for the Hindukush-Karakorum-Himalaya region. Hydrology and Earth System Science Discussion. 5, 865-902.

Akhtar, M., and Hussain, S.P., 2001. A study of climate change over Pakistan, MSc thesis, Postgraduate Center for Earth Sciences, University of the Punjab, Lahore, Pakistan.

Arakawa, A., and Lamb, V.R., 1977. Computational design of the basic dynamical process of the UCLA general circulation model. In: Methods in computational Physics, 17, Academic press, New York, 173-265.

Archer, D. R., 2003. Contrasting hydrological regimes in the Indus Basin. Journal of Hydrology. 274, 198-210.

Archer, D. R., 2004. Hydrological implications of spatial and altitudinal variation in temperature in the Upper Indus Basin. Nordic Hydrology. 35 (3), 209-222.

Archer, D. R., and Fowler, H. J., 2004. Spatial and temporal variations in precipitation in the Upper Indus Basin, global teleconnections and hydrological implications. Hydrology and Earth System Science. 8, 47-61.

Arnell, N. W., 1998, Climate Change and Water Resources in Britain. Climatic Change. 39, 83-110.

Arnell, N. W., 2003. Relative effects of multi-decadal climatic variability and changes in the mean and variability of climate due to global warming: future streamflow in Britain. Journal of Hydrology. 270, 195-213.

References

113

Arnell, N.W., 1999. A simple water balance model for the simulation of stream flow over a lare geographical domain. Journal of Hydrology. 217, 314-335.

Arnell, N.W., 2004. Climate change and global water resources: SRES emissions and socio-economic scenarios. Global Environmental Change. 14, 31-52.

Arnell, N.W., and Reynard, N.S., 1996. The effects of climate change due to global warming on river flows in Great Britain. Journal of Hydrology. 183, 397-424.

Bamzai, A. S., and Shukla, J., 1999. Relation between Eurasian Snow Cover, Snow Depth, and the Indian Summer Monsoon: An Observational Study. Journal of Climate. 12, 3117-3132.

Barnett, T. Malone, R. Pennell, W. Stammer, D. Semtner, B. and Washington, W. 2004. The Effects of Climate Change on Water Resources in the West: Introduction and Overview. Climatic Change. 62, 1-11.

Bergström, S., 1995. The HBV model, Chapter 13 of Computer models of watershed hydrology. Water Resource Publications. 443-476.

Bergström, S., Carlsson, B., Gardelin, M., Lindström, G., Pettersson, A., and Rummukainen, M., 2001. Climate change impacts on the runoff in Sweden-assessments by global climate models, dynamical downscaling and hydrological modelling. Climate Research. 16, 101-112.

Beven, K. and Binley, A., 1992. The future of distributed models: model calibration and uncertainty prediction. Hydrological Processes. 6, 279-298.

Bhutiyani, M. R., Kale, V. S., and Pawar, N. J., 2007. Long-term trends in maximum, minimum and mean annual air temperatures across the Northwestern Himalaya during the twentieth century. Climatic Change. 85, 159-177.

Bonan, G. B., 1998. The Land Surface Climatology of the NCAR Land Surface Model Coupled to the NCAR Community Climate Model. Journal of Climate. 11, 1307-1326.

Booij, M. J., 2002. Appropriate modelling of climate change impacts on river flooding, PhD thesis, University of Twente, the Netherlands, 206pp.

Booij, M. J., 2005. Impact of climate change on river flooding assessed with different spatial model resolutions. Journal of Hydrology. 303, 176-198.

Boorman, D. B., and Sefton, C. E., 1997. Recognizing the uncertainty in the quantification of the effects of climate change on hydrological response. Climatic Change. 35, 415-434.

Bowling, L. C., Lettenmaier, D. P., Nijssen, B., Graham, L. P., Clark, D. B., El Maayar, M., Essery, R., Goers, S., Gusev, Y. M., Habets, F., van den Hurk, B., Jin, J. M.,

References

114

Kahan, D., Lohmann, D., Ma, X.Y., Mahanama, S., Mocko, D., Nasonova, O., Niu, G. Y., Samuelsson, P., Shmakin, A. B., Takata, K., Verseghy, D., Viterbo, P., Xia, Y. L., Xue, Y. K., Yang, Z. L., 2003. Simulation of high latitude hydrological processes in the Torne-Kalix basin: PILPS Phase 2(e) 1: Experiment description and summary intercomparisons. Global and Planetary Change. 38, 1-30.

Burger, G., 1996. Expanded downscaling for generating local weather scenarios. Climate Research. 7, 111-128.

Burger, G., and Chen, Y., 2005. Regression-based downscaling of spatial variability for hydrologic applications. Journal of Hydrology. 311, 299-317.

Carr, A. T., 2003. Hydrologic comparisons and model Simulations of subarctic watersheds containing continuous and discontinuous permafrost, Seward Peninsula, Alaska, M.Sc thesis, University of Alaska Fairbanks, Alaska, 123pp.

Christensen, J. H., and Christensen, O. B., 2001. Regional Climate Simulations-A Study on Precipitation, Jørgensen, A., Fenger, J., and Halsnaes, K. (eds.), Climate Change Research: Danish Contributions, Danish Meteorological Institute/Danish Climate Centre, Copenhagen, pp. 151-166.

Christensen, J. H., and Christensen, O. B., 2003. Climate modelling: Severe summertime flooding in Europe. Nature. 421, 805-806.

Ciret, C. and Sellers, A.H., 1998. Sensitivity of ecosystem models to the spatial resolution of the NCAR Community Climate Model CCM2. Climate Dynamics. 14, 409-429.

Collins, W.J., Stevenson, D.S., Johnson, C.E., and Derwent, R.G., 1997. Tropospheric ozone in a global scale three dimensional Lagrangian model and its response to NOx emission control. Journal of Atmospheric Chemistry. 26, 223-274.

Cox, P.M., Betts, R.A., Bunton, C.B., Essery, R.L.H., Rowntree, P.R., and Smith, J., 1999. The impact of new land surface physics on the GCM simulation of climate and climate sensitivity. Climate Dynamics. 15, 183-203.

Cullen, M.J.P., 1993. The unified forecast/climate model. The Meteorological Magazine. 122, 81-94.

De Wit, M.J.M., Van den Hurk, B., Warmerdam, P.M.M., Torfs, P.J.J.F., Roulin, E., and Van Deursen, W.P.A., 2007. Impact of climate change on low-flows in the river Meuse. Climatic Change. 82, 351-372.

Deque, M., Jones, R.G., Wild, M., Giorgi, F., Christensen, J.H., Hassell, D.C., Vidale, P.L., Rockel, B., Jacob, D., Kjellstrom, E., de Castro, M., Kucharski, F., van den Hurk, B., 2005. Global high resolution versus Limited Area Model climate

References

115

change projections over Europe: quantifying confidence level from PRUDENCE results. Climate Dynamics. 25, 653-670.

Deque, M., Marquet, P., and Jones, R.G., 1998. Simulation of climate change over Europe using a global variable resolution general circulation model. Climate Dynamics. 14, 173-189.

Diaz-Nieto, J., and Wilby, R.L., 2005. A comparison of statistical downscaling and climate change factor methods: Impacts on low flows in the River Thames, United Kingdom. Climatic Change. 69, 245-268.

Dibike, Y. B., and Coulibaly, P., 2005. Hydrologic impact of climate change in the Saguenay watershed: comparison of downscaling methods and hydrologic models. Journal of Hydrology. 307, 145-163.

Dickinson, R. E., Errico, R. M., Giorgi, F., and Bates, G. T., 1989. A Regional Climate Model for Western United States. Climatic Change. 15, 383-422.

Diermanse, F.L.M., 2001. Physically based modelling of rainfall-runoff processes. PhD thesis. Delft University Press, Delft, 234pp.

Diolaiuti, G., Pecci, M., and Smiraglia, C., 2003. Liligo Glacier, Karakoram, Pakistan: a reconstruction of the recent history of a surge-type glacier. Annals of Glaciology. 36, 168-172.

Doll, P., 2002. Impact of climate change and variability on irrigation requirements: A global perspective. Climatic Change. 54, 269-293.

Downing, T.E., Butterfield, R.E., Bindi, M., Brooks, R.J., Carter, T.R., Delecolle, R., Harnos, Z.S., Harrison, P.A., Iglesias, A., New, M., Moss, S., Olesen, J.E., Orr, J.L., Porter, J., Semenov, M.A., and Wolf, J., 2002. Scaling Methods in Regional Integrated Assessments: From Points Upward and from Global Models Downwards. Integrated Assessment. 3, 167-187.

Durman, C.F., Gregory, J.M., Hassell, D.C., Jones, R.G., and Murphy, J.M., 2001 A comparison of extreme European daily precipitation simulated by a global and a regional climate model for present and future climates. Quarterly Journal of the Royal Meteorological Society. 127, 1005-1015.

Easterling, D. R., Horton, B., Jones, P. D., Peterson, T. C., Karl, T. R., Parker, D. E., Salinger, M. J., Razuvayev, V., Plummer, N., Jamason, P., and Folland, C.K., 1997. Maximum and minimum temperature trends for the globe. Science. 277, 364-367.

Eckhardt, K., and Ulbrich, U., 2003. Potential impacts of climate change on groundwater recharge and streamflow in a central European low mountain range. Journal of Hydrology. 284, 244-252.

References

116

Emori, S., Nozawa, T., Abe-Ouchi, A., Numaguti, A., and Nakajima, T., 1999. Coupled ocean-atmosphere model experiments of future climate change with an explicit representation of sulfate aerosol scattering. Journal of the Meteorological Society of Japan. 77, 1299-1307.

Essery, R. L. H., Best, M. J., Betts, R. A., Cox, P. M., and Taylor, C. M., 2003: Explicit representation of subgrid heterogeneity in a GCM land surface scheme. Journal of Hydrometeorology. 4, 530-543.

Flato, G. M., Boer, G.J., Lee, W. G., McFarlane, N. A., Ramsden, D., Reader, C., and Weaver, A. J., 2000. The Canadian Centre for Climate Modelling and Analysis global coupled model and its climate. Climate Dynamics. 16, 451-467.

Folland, C. K., Rayner, N. A., S. J. Brown, Smith, T. M., Shen, S. S. P., Parker, D. E., Macadam, I., Jones, P. D., Jones, R. N., Nicholls, N., and Sexton, D. M. H., 2001. Global temperature change and its uncertainties since 1861. Geophysical Research Letters. 28 (13), 2621-2624.

Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W., Haywood, J., Lean, J., Lowe, D.C., Myhre, J., Nganga, J., Prinn, R., Raga, G., Schulz, M., and Van Dorland, R., 2007. Changes in Atmospheric Constituents and Radiative Forcing. In: Climate Change 2007: The Physical Science Basis. Contribution f Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and Miller, H.L., (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Fowler, A. M., and Hennessy, K. J., 1995. Potential impacts of global warming on the frequency and magnitude of heavy precipitation. Natural Hazards. 11, 283-303.

Fowler, H. J., and Archer, D. R., 2005. Hydro-climatological variability in the Upper Indus Basin and implications for water resources. Regional Hydrological Impacts of Climatic Change-Impact Assessment and Decision Making, T. Wagener et al., Eds., IAHS Publication. 295, 131-138.

Fowler, H. J., and Archer, D. R., 2006: Conflicting signals of climate change in the Upper Indus Basin. Journal of Climate. 19, 4276-4293.

Fowler, H. J., Blenkinsop, S., and Tebaldi, C., 2007a. Linking climate change modelling to impacts studies: recent advances in downscaling techniques for hydrological modelling. International Journal of Climatology. 27, 1547-1578.

Fowler, H. J., and Kilsby, C. G., 2007b. Using regional climate model data to simulate historical and future river flows in northwest England. Climatic Change. 80, 337-367.

Fowler, H. J., Ekstrom, M., Kilsby, C.G., and Jones, P.D., 2005. New estimates of future changes in extreme rainfall across the UK using regional climate model

References

117

integrations. 1. Assessment of control climate. Journal of Hydrology. 300, 212-233.

Fowler, H. J., Kilsby, C.G., and Stunell, J., 2007b. Modelling the impact of projected climate change on water resources in north-west England. Hydrology and Earth System Science. 11, 1115-1126.

Frei, C., Christensen, J.H., Deque, M., Jacob, D., Jones, R.G., and Vidale, P.L., 2003. Daily precipitation statistics in regional climate models: evaluation and intercomparison for the European Alps. Journal of Geophysical Research. 108(D3), 4124-4142.

Frei, C., Scholl, R., Fukutome, S., Schmidli, J., and Vidale, P.L., 2006. Future change of precipitation extremes in Europe: an intercomparison of scenarios from regional climate models. Journal of Geophysical Research. 111, D06105, DOI: 10.1029/2005JD005965.

Gaffin, S., Xing, X., Yetman, G., 2004. Downscaling and geo-spatial gridding of socio-economic projections from the IPCC Special Report on Emissions Scenarios (SRES). Global Environmental Change. 14, 105-123.

Gellens, D. and Roulin, F., 1998, Streamflow Response of Belgian Catchments to IPCC Climate Change Scenarios. Journal of Hydrology. 210, 242-258.

Giorgi, F. and Mearns, L., 1991. Approaches to the simulation of regional climate change: a review. Review of Geophysics. 29, 191-216.

Giorgi, F., 1990. Simulation of regional climate using a limited area model nested in a general circulation model. Journal of Climate. 3, 941-963.

Giorgi, F., 2002. Dependence of the surface climate interannual variability on spatial scale. Geophysical Research Letters. 29, 2101, doi:10.1029/2002GL016175.

Giorgi, F., and Marinucci, M. R., 1996. An investigation of the sensitivity of simulated precipitation to model resolution and its implications of climate studies. Monthly Weather Review. 124, 148-166.

Giorgi, F., and Mearns, L.O., 1999. Regional climate modelling revisited. An introduction to the special issue. Journal of Geophysical Research. 104, 6335-6352.

Giorgi, F., Bi, X., and Pal, J. S., 2004. Mean, interannual variability and trends in a regional climate change experiment over Europe. I. Present-day climate (1961-1990). Climate Dynamics. 22, 733-756.

Giorgi, F., Brodeur, C. S., and Bates, G. T., 1994. Regional climate change scenarios over the United States produced with a nested regional climate model. Journal of Climate. 7, 375-399.

References

118

Gordon, C., Cooper, C., Senior, C. A., Banks, H., Gregory, J.M., Johns, T.C., Mitchell, J.F.B., Woods, R.A., 2000. The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Climate Dynamics. 16, 147-168

Graham, L. P., Andreasson, J., Carlsson, B., 2007. Assessing climate change impacts on hydrology from an ensemble of regional climate models, model scales and linking methods-a case study on the Lule River basin. Climatic Change. 81, 293-307.

Graham, P. L., Hagemann, S., Jaun, S., and Beniston, M., 2007b. On interpreting hydrological change from regional climate models. Climatic Change. 81, 97-122.

Grotch, S.L., and MacCracken, M.C., 1991. The use of general circulation models to predict regional climate change. Journal of Climate. 4, 286-303.

Guo, S., Wang, J., Xiong, L., Ying, A., and Li, D., 2002. A macro-scale and semi-distributed monthly water balance model to predict climate change impacts in China. Journal of Hydrology. 268, 1-15.

Gupta, H.V., Sorooshian, S. and Yapo, P.O., 1998. Toward improved calibration of hydrologic models: Multiple and noncommensurable measures of information. Water Resources Research. 34, 751-763.

Gutowski, W. J., Decker, S. G., Donavon, R. A., Pan, Z., Arritt, R. W., and Takle, E. S., 2003. Temporal-spatial scales of observed and simulated precipitation in central U.S. climate. Journal of Climate. 16, 3841-3847.

Haeberli, W., Frauenfelder, R., Hoelzle, M., Maisch, M., 1999. On rates and acceleration trends of global glacier mass changes. Geografiska Annaler. 81, 585-591.

Hagg, W., Braun, L.N., Kuhn, M., Nesgaard, T.I., 2007. Modelling of hydrological response to climate change in glacierized Central Asian catchments. Journal of Hydrology. 332, 40-53.

Hay, L.E., and Clark, M.P., 2003. Use of statistically and dynamically downscaled atmospheric model output for hydrologic simulations in three mountainous basins in the western United States. Journal of Hydrology. 282, 56-75.

Hay, L.E., Clark, M. P., Wilby, R.L., Gutowski, W.J., Leavesley, G. H., Pan, Z., Arritt, R.W., and Takle, E.S., 2002. Use of regional climate model output for hydrological simulations. Journal of Hydrometeorology. 3, 571-590.

Hellström, C. and Chen, D., 2003. Statistical downscaling based on dynamically downscaled predictors: application to monthly precipitation in Sweeden. Advances in Atmospheric Sciences. 20 , 951-958.

References

119

Hewitt, K. 2005. The Karakoram anomaly? Glacier expansion and the elevation effect, Karakoram Himalaya. Mountain Research and Development. 25, 332-340.

Hewitt, K. 2007. Tributary glacier surges: an exceptional concentration at Panmah Glacier, Karakoram Himalaya. Journal of Glaciology. 53, 181-188.

Hewitt, K., 1989: The altitudinal organization of Karakoram geomorphic processes and depositional environments. Z. Geomorph. N. F. Suppl.-Bd. 76, 9-32.

Hewitt, K., 1993. The altitudinal distribution of Karakoram geomorphic processes and depositional environments. In Shroder, J.F., Jr., ed. Himalaya to the Sea: geology, geomorphology and the Quaternary. New York, Routledge,159-183.

Hewitt, K., 1998. Glaciers receive a surge of attention in the Karakoram Himalaya. EOS, Transactions. American Geophysical Union. 79, 104-105.

Hewitt, K., 2003. The Karakoram Anomaly? Climate, Glacier Behaviour and the relief/elevation effect in the Karakoram Himalaya, Inner Asia, The Kenneth Hewitt Symposium, Cold Regions Research Centre, 25th April, Wilfrid Laurier University

Hingray, B., Mouhous, N., Mezghani, A., Bogner, K., Schaefli, B., Musy, A., 2007. Accounting for global-mean warming and scaling uncertainties in climate change impact studies: application to a regulated lake system. Hydrology and Earth System Sciences. 11, 1207-1226.

Hudson, D.A., Jones, R.G. 2002. Regional climate model simulations of present-day and future climates over southern Africa, Technical Note 39, Hadley Centre for Climate Prediction and Research, Met Office, Bracknell, U.K.

Huntingford, C., Jones, R.G., Prudhomme, C., Lamb, R., and Gash, J. H. C., 2003. Regional climate model predictions of extreme rainfall for a changing climate. Quarterly Journal of the Royal Meteorological Society. 129,1607-1621

Huntington, T.G., 2006. Evidence for intensification of the global water cycle: Review and synthesis. Journal of Hydrology. 319, 83-95.

IPCC, 2000. Special report on emission scenarios. Intergovernmental Panel on Climate Change [Nakicenovic, N., and Swart, R., (eds.), Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

IPCC, 2001. Climate Change 2001: The scientific basis, Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change [Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K., and Johnson, C.A., (eds.), Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 881pp.

References

120

IPCC, 2007. Climate Change 2007: The physical science basis, Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., Qin, D., Manning, M., Marquis, M., Averyt, K., Tignor, M.M.B., Miller, H.L., (eds.), Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Islam, M. N., Rafiuddin, M., Ahmed, A. U., and Kolli, R. K., 2008. Calibration of PRECIS in employing future scenarios in Bangladesh. International Journal of Climatology. 28, 617-628.

Jansson, P., Hock, R., and Schneider, T., 2003. The concept of glacier storage: a review. Journal of Hydrology. 282, 116-129.

Jasper, K., Calanca, P., Gyalistras, D. and Fuhrer, J., 2004. Differential impacts of climate change on the hydrology of two alpine river basins. Climate Research. 26, 113-129.

Johns, T.C., Carnell, R.E., Crossley, J.F., Gregory, J.M., Mitchell, J.F.B., Senior, C.A., Tett, S.F.B., and Wood, R.A. 1997. The second Hadley Centre coupled ocean-atmosphere GCM: Model description, spinup and validation. Climate Dynamics. 13, 103-134.

Jones, P. D., and Moberg, A., 2003. Hemispheric and large-scale surface air temperature variations: An extensive revision and an update to 2001. Journal of Climate. 16, 206-223.

Jones, P. D., M. New, D. E. Parker, S. Martin, and Rigor, I. G., 1999: Surface air temperature and its changes over the past 150 years. Reviews of Geophysics. 37, 173-199.

Jones, R.G., Noguer, M., Hassell, D.C., Hudson, D., Wilson, S.S., Jenkins, G.J., and Mitchell, J.F.B., 2004. Generating high resolution climate change scenarios using PRECIS, Me Office Hadley Centre, Exeter, UK, 40pp.

Joubert, A.M., and Tyson, P.D., 1996. Equilibrium and coupled GCM simulations of future southern African climates. South African Journal of Science. 92, 471-485.

Kahlown, M. A., Raoof, A., Zubair, M., and Kemper, W. D., 2007. Water use efficiency and economic feasibility of growing rice and wheat with sprinkler irrigation in the Indus Basin of Pakistan. Agriculture Water Management. 87, 292-298.

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, J., Zhu, Y., Leetmaa, A., Reynolds, B., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J., Mo, K. C., Ropelewski, C., Wang, J., Jenne, R., and Joseph, D., 1996. The NCEP/NCAR 40-Year Reanalysis Project. Bulletin of the American Meteorological Society. 77, 437-471.

References

121

Karl, T. R., Jones, P.D., Knight, R.W., Kukla, G., Plummer, N., Razuvayev, V., Gallo, K.P., Lindseay, J., Charlson, R.J. and Peterson, T.C., 1993. A new perspective on recent global warming: Asymmetric trends of daily maximum and minimum temperature. Bulletin of the American Meteorological Society. 74, 1007-1023.

Karl, T.R., Wang, W.C., Schlesinger, M.E., Knight, R.W. and Portman, D., 1990. A method of relating general circulation model simulated climate to observed local climate. Part I: Seasonal statistics. Journal of Climate. 3, 1053-1079.

Kay, A. L., Jones, R. G., and Reynard, N. S., 2006a. RCM rainfall for UK flood frequency estimation. I. Method and validation. Journal of Hydrology. 318, 151-162.

Kay, A. L., Jones, R. G., and Reynard, N. S., 2006b. RCM rainfall for UK flood frequency estimation. II. Climate change results. Journal of Hydrology. 318, 163-172.

Kharin, V. V., and Zwiers, F .W., 2000. Changes in the Extremes in an Ensemble of Transient Climate Simulations with a Coupled Atmosphere–Ocean GCM. Journal of Climate. 13, 3760-3788.

Killingtveit, A and Saelthun, N.R., 1995. Hydropower development: hydrology. Norwegian Inst. Technol., Osolo.

Kleinn, J., Frei, C., Gurtz, J., Luthi, D., Vidale, P. L., and Schar, C., 2005. Hydrologic simulations in the Rhine basin driven by a regional climate model. Journal of Geophysical Research. 110, D04102, Doi:10.1029/2004JD005143.

Klein-Tank, A. M. G., Peterson, T. C., Quadir, D. A, Dorji, S., Zou, X., Tang, H., Santhosh, K., Joshi, U. R., Jaswal, A. K., Kolli, R. K., Sikder, A. B., Deshpande, N. R., Revadekar, J. V., Yeleuova, K., Vandasheva, S., Faleyeva, M., Gomboluudev, P., Budhathoki, K. P., Hussain, A., Afzaal, M., Chandrapala, L., Anvar, H., Amanmurad, D., Asanova, V. S., Jones, P. D., New, M. G., and Spektorman, T., 2006. Changes in daily temperature and precipitation extremes in central and south Asia. Journal of Geophysical Research. 111, D16105, doi:10.1029/2005JD006316,

Kripalani1, R. H., Oh1, J. H., Kulkarni, A., Sabade, S. S., and Chaudhari, H. S., 2007. South Asian summer monsoon precipitation variability: Coupled climate model simulations and projections under IPCC AR4. Theoretical and Applied Climatology. 90, 133-159.

Krysanova, V., Bronstert, A., Muller-Wohlfeil, D. I., 1999. Modelling river discharge for large drainage basins: from lumped to distributed approach. Hydrological Science Journal. 44, 313-331.

References

122

Krysanova, V., Hattermann, F.F., and Wechsung, F., 2007. Implications of complexity and uncertainty for integrated modelling and impact assessment in river basins. Environmental Modelling and Software. 22, 701-709

Kumar, K. R., and Hingane, L. S., 1988. Long term variation of surface air temperature at major industrial cities of India. Climatic Change. 13, 287-307.

Kumar, R. K., Sahai, A.K., Kumar, K. K., Patwardhan, S. K., Mishra, P.K., Revadekar, J. V., Kamala, K., Pant, G.B., 2006. High-resolution climate change scenarios for India for the 21st century. Current Science. 90, 334-345.

Kundzewicz, Z.W., Mata, L.J., Arnell, N.W., Döll, P., Kabat, P., Jiménez, B., Miller, K.A., Oki, T., Sen, Z., and Shiklomanov, I.A., 2007. Freshwater resources and their management. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Parry, M.L., Canziani, O.F., Palutikof, J.P., van der Linden, P.J., and Hanson, C.E., Eds., Cambridge University Press, Cambridge, UK, 173-210.

Kunkel, K. E., Pielke, R. A., and Changnon, S. A., 1999. Temporal fluctuations in weather and climate extremes that cause economic and human health impacts: a review. Bulletin of the American Meteorological Society. 80, 1077-1098.

Kysely, J., 2002. Comparison of extremes in GCM-simulated, downscaled and observed central-European temperature series. Climate Research. 20, 211-222.

Leander, R., and Buishand, T.A., 2007. Resampling of regional climate model output for the simulation of extreme river flows. Journal of Hydrology. 332, 487-496.

Leggett J, Pepper W J, Swart R J. 1992. Emissions scenarios for the IPCC: an update.In: Climate change 1992: the supplementary report to the IPCC scientific assessment, Houghton J T, Callander B A, Varney S K (eds), 75-95.

Legutke, S., and Voss, R., 1999. The Hamburg atmosphere-ocean coupled circulation model ECHO-G. Technical report No. 18, German Climate Computre Centre (DKRZ), Hamburg, Germany, 62 pp

Lenderink, G., Buishand, A., and van Deursen, W., 2007. Estimates of future discharges of the river Rhine using two scenarios methodologies: direct versus delta approach. Hydrology and Earth System Sciences. 11, 1145-1159.

Leung, L. R., Qian, Y., Bian, X., 2003. Hydroclimate of the western United States based on observations and regional climate simulation of 1981-2000. Part I.:seasonal statistics. Journal of Climate. 16, 1892-1911.

Leung, L.R., Qian, Y., Bian, X., Washington, W.M., Han, J., and Roads, J.O., 2004. Mid-century ensemble regional climate change scenarios for the Western United States. Climatic Change. 62, 75-113.

References

123

Li, X. and Sailor, D., 2000. Application of tree structured regression for regional precipitation prediction using general circulation model output. Climate Research. 16, 17-30.

Liden, R., and Harlin, J., 2000. Analysis of conceptual rainfall-runoff modelling performance in different climates. Journal of Hydrology. 238, 231-247.

Lijuan, C., Wenjie, D., Yinlong, X., Yong, Z., and Michael, S., 2007. Validating the Runoff from the PRECIS Model Using a Large-Scale Routing Model. Advances in Atmospheric Research. 24, 855-862.

Linderson, M.L, Achberger, C. and Chen, D., 2004. Statistical downscaling and scenario construction of precipitation in Scania, southern Sweden. Nordic Hydrology. 35, 261-287.

Lindström, G., Johansson, B., Persson, M., Gardelin, M. and Bergström, S., 1997. Development and test of the distributed HBV-96 model. Journal of Hydrology. 201, 272-288.

Loaiciga, H.A., Valdes, J.B., Vogel, R., and Schwarz, H., 1996. Global warming and the hydrologic cycle. Journal of Hydrology. 174, 83-127.

Loukas, A., Vasiliades, L. and Dalezios, N.R., 2002. Potential climate change impacts on flood producing mechanisms in southern British Columbia, Canada using the CGCMA1 simulation results. Journal of Hydrology. 259, 163-188.

Loukas, A., Vasiliades, L., and Dalezios, N. R., 2004. Climate change implications on flood response of a mountainous watershed. Water Air and Soil Pollution: Focus. 4, 331-347.

Mann, M. E., and Jones, P. D., 2003. Global surface temperatures over the past two millennia. Geophysical Research Letters. 30 (15), 1820, doi:10.1029/2003GL017814.

Mastny, L., 2000. Melting of earth’s ice cover reaches new high. World Watch Institute,Washington,DC. [http://www.worldwatch.org/alerts/000306.html]

May, W. and Roeckner, E., 2001. A Time-Slice Experiment with the ECHAM 4 AGCM at High Resolution: The Impact of Resolution on the Assessment of Anthropogenic Climate Change. Climate Dynamics. 17, 407-420.

Mayer, C., Lambrecht, A., Belo, M., Smiraglia, C., Diolaiuti, G., 2006. Glaciological characteristics of the ablation zone of Baltoro glacier, Karakoram, Pakistan. Annals of Glaciology. 43, 123-131.

McGregor, J.L., 1997. Regional climate modeling. Meteorology and Atmospheric Physics. 63, 105-117.

References

124

Meehl, G.A., Stocker, T. F., Collins, W D., Friedlingstein, P., Gaye, A. T., Gregory, J M., Kitoh, A., Knutti, R., Murphy, J M., Noda, A., Raper, S.C.B., Watterson, I.G., Weaver, A.J. and Zhao, Z. C., 2007. Global climate projections In: Climate Change 2007: The Physical Science Basis. Contribution f Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and Miller, H.L., (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Menzel, L., and Bürger, G., 2002. Climate change scenarios and runoff response in the Mulde catchment (Southern Elbe, Germany). Journal of Hydrology. 267, 53-64.

Menzel, L., Thieken, A. H., Schwandt, D., Bürger, G., 2006. Impact of climate change on the regional Hydrology-Scenarios-Based modelling studies in the German Rhine Catchment. Natural Hazards. 38, 45-61.

Middelkoop, H., Daamen, K., Gellens, D., Grabs, W., Kwadijk, J.C.J., Lang, H., Parmet, B.W.A.H., Schädler, B., Schulla, J., and Wilke, K., 2001. Impact of Climate Change on Hydrological Regimes and Water Resources Management in the Rhine Basin. Climatic Change. 49, 105-128.

Milly, P. C.D., Wetherald, R. T., Dunne, K. A. and Delworth, T. L., 2002. Increasing risk of great floods in a changing climate. Nature. 415, 514-517.

Mitchell T. D., and Jones., P.D., 2005. An improved method of constructing a database of monthly climate observations and associated high-resolution grids. International Journal of Climatology. 25, 693-712.

Mohseni, O. and Stefan, H.G., 2001. Water budgets of two watersheds in different climatic zones under projected climate warming. Climatic Change. 49, 77-104.

Nakicenovic, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gaffin, S., Gregory, K., Grübler, A., Jung, T. Y., Kram, T., La Rovere, E. L., Michaelis, L., Mori, S., Morita, T., Pepper, W., Pitcher, H., Price, L., Riahi, K., Roehrl, A., Rogner, H-H., Sankovski, A., Schlesinger, M., Shukla, P., Smith, S., Swart, R., van Rooijen, R., Victor, N., Dadi, Z., 2000. IPCC special report on emission scenarios, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Nash, J. E., and Sutcliffe, J. V., 1970. River flow forecasting through conceptual models 1: a discussion of principles. Journal of Hydrology. 10, 282-290.

NESPAK, 1997. Flood early warning system manual for Indus Basin, National Engineering Services Pakistan (NESPAK), Lahore, pp 425.

New, M., Hulme, M., and Jones, P., 2000. Representing Twentieth-Century Space–Time Climate Variability. Part II: Development of 1901–96 Monthly Grids of Terrestrial Surface Climate. Journal of Climate. 13, 2217-2238.

References

125

New, M., Todd, M., Hulme, M., Jones, P., 2001. Precipitation measurements and trends in the twentieth century. International Journal of Climatology. 21, 1889-1922.

Nijssen, B., Bowling, L. C., Lettenmaier, D. P., Clark, D. B., El Maayar, M., Essery, R., Goers, S., Gusev, Y. M., Habets, F., van den Hurk, B., Jin, J. M., Kahan, D., Lohmann, D., Ma, X.Y., Mahanama, S., Mocko, D., Nasonova, O., Niu, G. Y., Samuelsson, P., Shmakin, A. B., Takata, K., Verseghy, D., Viterbo, P., Xia, Y. L., Xue, Y. K., Yang, Z. L., 2003. Simulation of high latitude hydrological processes in the Torne-Kalix basin: PILPS Phase 2(e) 2: Comparison of model results with observations. Global and Planetary Change. 38, 31-53.

Noguer, M., Jones, R. G., and Murphy, J., 1998. Sources of systematic errors in the climatology of a nested Regional Climate Model (RCM) over Europe. Climate Dynamics. 14, 691-712.

Oerlemans, J., Anderson, B., Hubbard, A., Huybrechts, Ph., Johannesson, T., Knap, W.H., Schmeits, M., Stroeven, A.P., van de Wal, R.S.W., Wallinga, J., Zuo, Z., 1998. Modelling the response of glaciers to climate warming. Climate Dynamics. 14, 267-274.

P & D, 2001. A report of planning and Development, Government of Pakistan.

Paul, F., Kaab, A., Maisch, M., Kellenberger. T., Haeberli, W., 2004. Rapid disintegration of Alpine glaciers observed with satellite data. Geophysical Research Letters. 31, L21402, doi:10.1029/2004GL020816.

Paul, F., Maisch, M., Rothenbuhler, C., Hoelzle, M., Haeberli, W., 2007. Calculation and visualisation of future glacier extent in the Swiss Alps by means of hypsographic modelling. Global and Planetary Change. 55, 343-357.

PCO. 2008. Population clock. Available at http://www.statpak.gov.pk/depts/pco/index.html [cited 2008: verified 01 March 2008]. Population Census Organization.

Perks, L.A., Schulze, R.E., Kiker, G.A., Horan, M.J.C., and Maharaj, M., 2000. Preparation of climate data and information for application in impact studies of climate change over southern Africa, Report to South African Country Studies for Climate Change Programme. ACRUcons Report. 33, pp74.

Peterson, B. J., Holmes, R. M., McClelland, J. W., Vörösmarty, C. J., Lammers, R. B., Shiklomanov, A. I., Shiklomanov, I. A., and Rahmstorf, S., 2002. Increasing River Discharge to the Arctic Ocean. Science. 298, 2171-2173.

Pilling, C. G., and Jones, J. A. A., 2002. The impact of future climate change on seasonal discharge, hydrological processes and extreme flows in the Upper Wye experimental catchment, mid-Wales. Hydrological Processes. 16, 1201-1213.

References

126

Piltz, R. (Ed.), 1998. Our Changing Planet: The FY 1998 U.S. Global Research Program. A Report to the Subcommittee on Global Change Research, Committee on Environment and Natural Resources of the National Science and Technology Council. A Supplement to the President, a Fiscal Year 1998 Budget.

Prudhomme, C., Reynard, N., Crooks, S., 2002. Downscaling of global climate models for flood frequency analysis: Where are we now?. Hydrological Processes. 16, 1137-1150.

Raicich, F., Pinardi, N., and Navarra, A., 2003. Teleconnections between Indian monsoon and Sahel rainfall and the Mediterranean. International Journal of Climatology. 23, 173-186.

Raisanen, J., 2007. How reliable are climate models?. Tellus. 59A, 2-29.

Rango, A., 1992. Worldwide testing of the snowmelt runoff model with applications for predicting the effects of climate change. Nordic Hydrology. 23, 155-172.

Refsgaard, J.C. and Storm, B., 1996. Construction, calibration and validation of hydrological models. In: J.C. Refsgaard (Editor), Distributed hydrological modelling. Kluwer Academic Press, Dordrecht, Netherlands, pp. 41-54.

Rosenberg, N. J., Brown, R. A., Izaurralde, R. C., and Thomson, A. M., 2003. Integrated assessment of Hadley Centre (HadCM2) climate change projections on agricultural productivity and irrigation water supply in the conterminous United States I. Climate change scenarios and impacts on irrigation water supply simulated with the HUMUS model. Agricultural and Forest Meteorology. 117, 73-96.

Saier Jr, M.H., 2007. Climate Change, 2007. Water Air Soil Pollution. 181, 1-2.

Salathe, E. P., 2003. Comparison of various precipitation downscaling methods for the simulation of streamflow in a rainshadow river basin. International Journal of Climatology. 23, 887-901.

Santer, B.D., 1996. A search for human influences on the thermal structure of the atmosphere. Nature. 384, 39-46.

Santer, B.D., Bruggemann, W., Cubasch, U., Hasselmann, K., Hock, H., Maier Reimer, E., and Mikolajewicz, U., 1994. Signal-to-noise analysis of time-dependent greenhouse warming experiments. Part I: Pattern analysis. Climate Dynamics. 9, 267-285.

Santer, B.D., Taylor, K.E., Wigley, T.M.L., Penner, J.E., Jones, P.D. and Cubasch, U., 1995. Towards the detection and attribution of an anthropogenic effect on climate. Climate Dynamics. 12, 77-100.

References

127

Schafli, B., 2005. Quantification of modelling uncertainties in climate change impact studies on water resources: application to a glacier-fed hydropower production system in the Swiss Alps, PhD thesis, Lausanne, EPFL.

Schmidt, G.A., Ruedy, R., Hansen, J.E., Aleinov, I., Bell, N., Bauer, M., Bauer, S., Cairns, B., Canuto, V., Cheng, Y., Del Genio, A., Faluvegi, G., Friend, A.D., Hall, T.M., Hu, Y., Kelley, M., Kiang, N.Y., Koch, D., Lacis, A.A., Lerner, J., Lo, K.K., Miller, R.L., Nazarenko, L., Oinas, V., Ja. Perlwitz, Ju. Perlwitz, Rind, D., Romanou, A., Russell, G.L., Mki. Sato, Shindell, D.T., Stone, P.H., Sun, S., Tausnev, N., Thresher, D., and Yao, M.S., 2006. Present day atmospheric simulations using GISS ModelE: Comparison to in-situ, satellite and reanalysis data. Journal of Climate. 19, 153-192.

Seeber, L. and Gornitz, V., 1983. River profiles along the Himalayan arc as indicators of active tectonics. Tectonophysics. 92, 335-367.

Shrestha, M. L., and Shrestha, A. B., 2004. Recent trends and potential climate change impacts on glacier retreat/glacier lakes in Nepal and potential adaptation measures. OECD Global Forum on Sustainable Development: Development and Climate Change, Paris, France, OECD, 5-14.

SIHP, 1990. Snow and Ice Hydrology Project, Upper Indus Basin, Overall. Report, WAPDA- IDRC- Wilfrid Laurier University, 179pp.

Sillmann, J., and Roeckner, E., 2007. Indices for extreme events in projections of anthropogenic climate change. Climatic Change. 86, 83-104.

Singh, P., and Bengtsson, L., 2005. Impact of warmer climate on melt and evaporation for the rainfed, snowfed and glacierfed basins in the Himalayan region. Journal of Hydrology. 300, 140-154.

Singh, P., and Kumar, K., 1997. Effect of orography on precipitation in the western Himalayan region. Journal of Hydrology. 199, 183-206.

SMHI, 2005. Integrated hydrological modelling system (IHMS), Manual version 5.6, Swedish Meteorological and Hydrological Institute, Norrkoping.

Solman, S. A., Nunez, M. N., and Cabre, M. F., 2008. Regional climate change experiments over southern South America. I: present climate. Climate Dynamics. 30, 533-552.

Stott, P. A., Tett, S. F. B., Jones, G. J., Allen, M. R., Mitchell, J. F.B. and Jenkins, G. .J., 2000. External control of 20th century temperature by natural and anthropogenic forcings. Science. 290, 2133-2137.

Syed, F. S., Giorgi, F., Pal, J. S., King, M. P., 2006. Effect of remote forcings on the winter precipitation of central southwest Asia part 1:observations. Theoretical and Applied Climatology. 86, 147-160.

References

128

Te Linde, A.H., Aerts, J.C.J.H., Hurkmans, R.T.W.L., Eberle, M., 2007. Comparing model performance of two rainfall-runoff models in the Rhine basin using different atmospheric forcing data sets. Hydrololgy and Earth System Sciences Discussion. 4, 4325-4360.

Tett, S.F.B., Betts, R., Crowley, T.J., Gregory, J., Johns, T.C., Jones, A., Osborn, T.J., Ostrom, E., Roberts, D.L., and Woodage, M.J., 2007. The Impact of natural and anthropogenic forcings on climate and hydrology since 1550. Climate Dynamics. 28, 3-34.

Thorne, P. W., Jones, P.D., Tett, S.F.B., Allen, M.R., Parker, D.E., Stott, P.A., Jones, G.S., Osborn, T.J. and Davies, T.D., 2003. Probable causes of late twentieth century tropospheric temperature trends. Climate Dynamics. 21, 573-591.

Trenberth K.E., Jones, P.D., Ambenje, P., Bojariu, R., Easterling, D., Klein Tank, A., Parker, D., Rahimzadeh, F., Renwick, J.A., Rusticucci, M., Soden, B., and Zhai, P., 2007. Observations: Surface and Atmospheric Climate Change. In: Climate Change 2007: The Physical Science Basis. Contribution f Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and Miller H.L., (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Trenberth, K. E., 1998. Atmospheric moisture residence times and cycling: implications for rainfall rates and climate change. Climatic Change. 39, 667-694.

Trenberth, K. E., 1999. Conceptual Framework for Changes of Extremes of the Hydrological Cycle with Climate Change. Climatic Change. 42, 327-339.

Trenberth, K. E., and Shea, D. J., 2005: Relationship between precipitation and surface temperature. Geophysical Research Letters. 32, 10.1029/2005gl022760.

Trenberth, K. E., Dai, A., Rasmussen, R. M., and Parsons, D. B., 2003: The changing character of precipitation. Bulletin of the American Meteorological Society. 84, 1205-1217.

Treydte, K. S., Schleser, G. H., Helle, G., Frank, D. C., Winiger, M., Haugh, G. H., and Esper, J., 2006. The twentieth century was the wettest period in northern Pakistan over the past millennium. Nature. 440, 1179-1182.

Tripathi, S., Srinivas, V.V., and Nanjundiah, R. S., 2006. Downscaling of precipitation for climate change scenarios: A support vector machine approach. Journal of Hydrology. 330, 621-640.

Uchiyama, T., Mizuta, R., Kamiguchi, K., Kitoh, A., and Noda, A., 2006. Changes in Temperature-based Extremes Indices Due to Global Warming Projected by a Global 20-km-mesh Atmospheric Model. SOLA. 2, 68-71.

References

129

Uhlenbrook, S., Seibert, J., Leibundgut, C., Rodhe, A., 1999. Prediction uncertainty of conceptual rainfall-runoff models caused by problems in identifying model parameters and structure. Hydrological Science Journal. 44, 779-797.

Uppala, S.M., Kållberg, P.W., Simmons, A.J., Andrae, U.,Da Costa Bechtold, V., Fiorino, M., Gibson, J.K., Haseler, J., Hernandez, A., Kelly, G.A., Li, X., Onogi, K., Saarinen, S., Sokka, N., Allan, R.P., Andersson, E., Arpe, K., Balmaseda, M.A., Beljaars, A.C.M., Van De Berg, L., Bidlot, J., Bormann, N., Caires, S., Chevallier, F., Dethof, A., Dragosavac, M., Fisher, M., Fuentes, M., Hagemann, S., Holm, E., Hoskins, B.J., Isaksen, L., Janssen, P.A.E.M., Jenne, R., McNally, A.P., Mahfouf, J.F., Morcrette, J.J., Rayner, N.A., Saunders, R.W., Simon, P., Sterl, A., Trenberth, K.E., Untch, A., Vasiljevic, D., Viterbo, P., Woollen, J., 2005. The ERA-40 re-analysis. Quarterly Journal of the Royal Meteorological Society. 131, 2961-3012.

van Vuuren, D.P., Lucas, P.L. and Hilderink, H., 2007. Downscaling drivers of global environmental change: Enabling use of global SRES scenarios at the national and grid levels. Global Environmental Change. 17, 114-130.

Vavrus, S., 2007. The role of terrestrial snow cover in the climate system. Climate Dynamics. 29, 73-88.

von Storch, H., Zorita, E. and Cubash, U., 1993. Downscaling of global climate change estimates to regional scales: an application to Iberian rainfall in wintertime. Journal of Climate. 6, 1161-1671.

Wake, C. P., 1987. Snow accumulation studies in the central Karakoram. Proc. Eastern Snow Conf. 44th Annual Meeting Frederiction, Canada. 19-33.

Wang, S., McGrath, R., Semmler, T., Sweeney, C., Nolan, P., 2006. The impact of the climate change on discharge of Suir river catchment (Ireland) under different climate scenarios. Natural Hazards Earth System Science. 6, 387-395.

Wang, Y., Leung, L.R., McGregor, J.L., Lee, D.K., Wang, W.C., Ding, Y., and Kimura, F., 2004. Regional climate modeling: progress, challenges, and prospects. Journal of the Meteorological Society of Japan. 82, 1599-1628.

WAPDA, 1990. Water and Power Development Authority, Water Sector Investment Planning Study (WSIPS), (Lahore, Federal Planning Cell), 5 vols.

Watson, R. T., Zinyowera, M. C., Moss, R. H., 1996. Climate Change 1995: Impacts, Adaptations, and Mitigations of climate change, Cambridge University Press, 889 pp.

Weerts, A.H., 2003. Assessing and quantifying the combined effect of model parameter and boundary uncertainties in model based flood forecasting. Geophysical Research Abstract. 5, 14564.

References

130

Wescoat, J. L. Jr., 1991. Managing the Indus River basin in light of climate change: four conceptual approaches. Global Environmental Change. 1, 381-395.

Wetterhall, F., Halldin, S., Xu, C.Y., 2007. Seasonality properties of four statistical-downscaling methods in central Sweden. Theoretical and Applied Climatology. 87, 123-137.

WGMS, 2002. World Glacier Monitoring Service, Glacier mass balance data 2000/01. [www.geo.unizh.ch/wgms/]

Wilby, R. L., and Harris, I., 2006. A framework for assessing uncertainties in climate change impacts: low flow scenarios for the River Thames, UK. Water Resources Research. 42: W02419, DOI: 10.1029/2005WR004065.

Wilby, R. L., Hay, L. E., and Leavesley, G. H., 1999. A comparison of downscaled and raw GCM output: Implications for climate change scenarios in the San Juan River Basin, Colorado. Journal of Hydrology. 225, 67-91.

Wilby, R.L, Charles, S.P., Zorita, E., Timbal, B., Whetton, P., and Mearns, L.O., 2004: Guidelines for Use of Climate Scenarios Developed from Statistical Downscaling Methods, IPCC Task Group on Data and Scenario Support for Impact and Climate Analysis, 27 pp.

Wilby, R.L. and Wigley, T.M.L., 1997. Downscaling General Circulation Model output: a review of methods and limitations. Progress in Physical Geography. 21, 530-548.

Willis, I. and Bonvin, J.-M., 1995. Climate change in mountain environments: hydrological and water resource implications. Geography. 80, 247-261.

Wilson, M.F., and Hendersen-Sellers, A., 1985. A global archive of land cover and soils data for use in general circulation models. Journal of Climatology. 5, 119-143.

Wilson, S., Hassell, D., Hein, D., Jones, R., Taylor, R., 2005. Installing and using the Hadley Centre regional climate modelling system, PRECIS (version 1.4), Met Office Hadley Centre, Exeter, UK.

Winger, M., Gumpert, M., and Yamout, H., 2005. Karakorum-Hindukush-western Himalaya: assessing high-altitude water resources. Hydrological Processes. 19, 2329-2338.

Wood, A. W., Leung, L. R., Sridhar, V., and Lettenmaier, D. P., 2004. Hydrologic implications of dynamical and statistical approaches to downscaling climate model outputs. Climatic Change. 62, 189-216.

Woth, K., 2006. Regionalization of global climate change scenarios: An ensemble study of possible changes in the North Sea storm surge statistics, PhD Thesis, GKSS research centre, Hamburg, Germany.

References

131

Xu, C. Y., 1999. Estimation of Parameters of a ConceptualWater Balance Model for Ungauged Catchments. Water Resources Management. 13, 353-368.

Xu, C. Y., and Singh, V. P., 2004. Review on Regional Water Resources Assessment Models under Stationary and Changing Climate. Water Resources Management. 18, 591-612.

Yadav, R. R., W.-K. Park, J. Singh, and Dubey, B., 2004. Do the western Himalayas defy global warming? Geophysical Research Letters. 31, L17201, doi:10.1029/2004GL020201.

Yinlong, X., Yong, Z., Erda, L., Wantao, L., Wenjie, D., Jones, R., Hassell, D., Wilson, S., 2006. Analysis on the climate change responses over China under SRES B2 scenario using PRECIS. Chinese Science Bulletin. 51, 2260-2267.

132

ANNEXURE

PUBLISHED SCIENTIFIC PAPERS

Akhtar, M., Ahmad, N. and Booij, M.J.2008. Use of regional climate model simulations as input for hydrological models for the Hindukush-Karakorum-Himalaya region. Hydrology and Earth System Science Discussion. 5, 865-902.

Akhtar, M., Ahmad, N., and Booij, M.J., 2008. The impact of climate change on the water resources of Hindukush-Karakorum-Himalaya region under different glacier coverage scenarios. Journal of Hydrology. 355, 148-163.

Akhtar, M., Ahmad, N, Chaudhry, M. N, and Babur, K. 2005. Spatial and temporal variations in precipitation and temperature in the Upper Indus Basin, Pakistan, Proceedings of International Conference Environmentally Sustainable Development (ESDev-2005), 25-27 June, 2005, COMSATS, Institute of Information Technology, Abbotabad- Pakistan, 639-651.

Akhtar, M., Ahmad, N, Chaudhry, M. N, and Hussain, S.P. 2005. Climate change in the Upper Indus Basin of Pakistan: A Case Study, proceedings of National Workshop on “ Global change prospective in Pakistan–challenges, impacts, opportunities and prospects” 28-30 April, 2005, Pakistan Academy of Science (PAS), Islamabad-Pakistan, edited by Dr Amir Muhammad and Dr. Sajidin Hussain, 14-28.

the climate change impact on water resources of...

Documents