The Climate Change Impact on Water Resources of Upper Indus Basin-Pakistan Institute of Geology University of the Punjab, Lahore

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

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, PhD Professor Institute of Geology University of the Punjab Lahore, Pakistan

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

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.

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.

iii

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- 39 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) Table 3.3 Biases in mean precipitation (%) as simulated with the PRECIS- 47 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) Table 3.4 Seasonal changes of mean temperature and precipitation under 54 SRES B2 scenario from PRECIS in 2071-2100 over the seven sub regions relative to 1961-1990 (Summer = April-September; Winter=October-March) Table 4.1 Characteristics of study area 58 Table 4.2 Temperature and precipitation during two monsoon events at 60 selected stations Table 4.3 Biases in mean temperature (˚C) as simulated with PRECIS RCMs 63 relative to CRU reference data for different seasons and river basins (Winter =October-March; Summer =April-September) Table 4.4 Biases in precipitation (%) as simulated with PRECIS RCMs 63 relative to CRU reference data for different seasons and river basins (Winter = October- March; Summer = April-September) Table 4.5 Values and range of important parameters found in different 75 studies using HBV model Table 4.6 Parameter values for HBV for three river basins with three 75 different input data sets Table 4.7 Performance of three HBV models during calibration and 76 validation periods in different river basins Table 4.8 Efficiency Y of three HBV models using data sources different 86 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 Table 5.1 Seasonal changes of mean temperature and precipitation under 94 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)

Table 5.2 Mean relative change in future discharge (2071-2100) in a 102 changed SRES B2 climate relative to the present discharge (1961- 1990) for three glaciations stages and for three river basins Table 5.3 Characteristics of future annual maximum discharge simulated by 106 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

LIST OF FIGURES

Figure 1.1 Indus river basin 4 Figure 1.2 Mean annual hydrograph at different locations at the Indus river 5 including some headwater tributaries Figure 3.1 Topography of selected domain (a) Topography of the global 25 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 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 29 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) Figure 3.4 Mean seal level pressure (MSLP) for the period 1981-1990 for 31 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 Figure 3.5 Observed and simulated (baseline) patterns of annual 34 temperature (˚C) for (a) CRU data, (b) HadAM3P, (c) PRECIS- Had and (d) PRECIS-ERA Figure 3.6 Bias of annual temperature (˚C) for (a) PRECIS-Had and (b) 36 PRECIS-ERA with respect to CRU data Figure 3.7 Observed and simulated (PRECIS-Had and HadAM3P) annual 37 cycle of temperature averaged over the seven sub regions of figure 3.3 Figure 3.8 Observed and simulated (PRECIS-ERA) annual cycle of 38 temperature averaged over the seven sub regions of figure 3.3 Figure 3.9 Observed (CRU) and simulated (PRECIS-Had and HadAM3P) 39 seasonal temperature standard deviation averaged over the seven subregions of figure 3.3 Figure 3.10 Observed and simulated (baseline) patterns of annual 42 precipitation (mm/day) for (a) CRU data, (b) HadAM3P, (c) PRECIS-Had and (d) PRECIS-ERA Figure 3.11 Bias of annual precipitation (mm/day) for (a) PRECIS-Had and 44 (b) PRECIS-ERA with respect to CRU data Figure 3.12 Observed and simulated (PRECIS-Had and HadAM3P) annual 45 cycle of precipitation averaged over the seven sub regions of figure 3.3

Figure 3.13 Observed and simulated (PRECIS-ERA) annual cycle of 46 precipitation averaged over the seven sub regions of figure 3.3 Figure 3.14 Observed (CRU) and simulated (PRECIS-Had and HadAM3P) 47 seasonal precipitation coefficient of variation (CV) averaged over the seven sub regions of figure 3.3 Figure 3.15 Observed and simulated wet day frequencies averaged over the 48 seven sub regions shown in figure 3.3. Figure 3.16 Changes of mean temperature under SRES B2 scenario relative 50 to present day climate (a) Annual, (b) Summer and (c) Winter Figure 3.17 Changes of mean precipitation under SRES B2 scenario relative 51 to present day climate (a) Annual, (b) Summer and (c) Winter Figure 3.18 Annual cycle of temperature averaged over the seven sub 52 regions for present (1961-1990) climate and future (2071-2100) climate under SRES B2 scenario Figure 3.19 Annual cycle of precipitation averaged over the seven sub 53 regions for present (1961-1990) climate and future (2071-2100) climate under SRES B2 scenario 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 61 the rainfall events of (a) October, 1987 and (b) August, 1997 Figure 4.3 Mean annual cycle of temperature [˚C] over (a) Hunza river 64 basin, (b) Gilgit river basin and (c) Astore river basin as simulated with PRECIS RCMs and from CRU data Figure 4.4 Mean annual cycle of precipitation [mm/day] over (a) Hunza 65 river basin, (b) Gilgit river basin and (c) Astore river basin as simulated with PRECIS RCMs and from CRU data Figure 4.5 A schematic diagram of the hydrological model HBV (modified 71 after Lindström et al., 1997), numbers in brackets refer to described equations Figure 4.6 Sensitivity of HBV model parameters for (a) Hunza river basin, 77 (b) Gilgit river basin and (c) Astore river basin Figure 4.7 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV- 78 Met (Hunza river basin), with Y as a function of FC, GMELT, TT and DTTM Figure 4.8 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV- 79 ERA (Hunza river basin), with Y as a function of FC, GMELT, TT and DTTM Figure 4.9 Sensitivity analysis with FC, GMELT, TT and DTTM for HBV- 80 PRECIS (Hunza river basin), with Y as a function of FC, GMELT, TT and DTTM

vii

Figure 4.10 Observed and simulated discharge (m3/s) of (a) HBV-Met, (b) 81 HBV-ERA and (c) HBV-PRECIS for Hunza river basins during calibration period Figure 4.11 Double mass-curve analysis relating observed and simulated 82 discharge (m3/s) of (a) HBV-Met, (b) HBV-ERA and (c) HBV- PRECIS for Hunza river basin during calibration period Figure 4.12 Observed and simulated (HBV-Met, HBV-PRECIS and HBV- 83 ERA) mean annual discharge (m3/s) cycle of (a) Hunza river basin, (b) Gilgit river basin and (c) Astore river basin Figure 4.13 Observed, HBV-Met simulated and HBV-PRECIS simulated 87 annual maximum discharge as a function of return period for three river basins in the present day climate Figure 4.14 Observed discharge (green line) and uncertainties in discharge 88 (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Hunza river basin during the 1986 hydrological year Figure 4.15 Observed discharge (green line) and uncertainties in discharge 89 (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Gilgit river basin during the 1986 hydrological year Figure 4.16 Observed discharge (green line) and uncertainties in discharge 90 (red shade) simulated through a) HBV-Met, b) HBV-ERA and c) HBV-PRECIS for Astore river basin during the 1986 hydrological year Figure 5.1 Mean annual cycle of temperature [˚C] over river basins (a) 95 Hunza, (b) Gilgit and (c) Astore as simulated with PRECIS for present (1961-90) and future (2071-2100) climate under SRES B2 scenario Figure 5.2 Mean annual cycle of precipitation [mm/day] over river basins 96 (a) Hunza, (b) Gilgit and (c) Astore as simulated with PRECIS for present (1961-90) and future (2071-2100) climate under SRES B2 scenario Figure 5.3 Annual discharge cycle simulated by HBV-Met for the present 100 climate and future SRES B2 climate for three stages of glaciation for three river basins Figure 5.4 Annual discharge cycle simulated by HBV-PRECIS for the 101 present climate and future SRES B2 climate for three stages of glaciation for three river basins Figure 5.5 HBV-PRECIS simulated annual maximum discharge as a 104 function of return period for current and changed SRES B2 climate for three glacier stages for three river basins Figure 5.6 HBV-Met simulated annual maximum discharge as a function of 105 return period for current and changed SRES B2 climate for three glacier stages for three river basins

viii

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 22 SCENARIOS

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 48 2071-2100

3.6 Summary 54

CHAPTER 4 PRECIS SIMULATIONS AS INPUT TO 57 HYDROLOGICAL MODELLING

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 93 RESOURCES

5.1 Background 93

5.2 Change of Temperature and Precipitation in the Selected River 93 Basins

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 108 FUTURE WORK

6.1 Conclusions 108

6.2 Recommendations for Future Research Work 111

REFERENCES 112

ANNEXURE LIST OF PUBLICATIONS 132

Chapter 1 Introduction

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.

1 Chapter 1 Introduction

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

2 Chapter 1 Introduction 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).

3 Chapter 1 Introduction

Figure 1.1 Indus River Basin

4 Chapter 1 Introduction

8000

Yogo Alam Bridge 6000 Beshame Doyian Gilgit Kachura /s)

3 Partab Shigar Dainyor 4000 Karmong Tarbela Discharge(m

2000

0 JAN MAR MAY JUL SEP NOV Time Period (Month)

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

5 Chapter 1 Introduction

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

Length Area Elevations (m) Glacier 2 (Km) (Km ) 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

6 Chapter 1 Introduction

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

7 Chapter 1 Introduction 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

8 Chapter 1 Introduction 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.

9 Chapter 1 Introduction

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.

10 Chapter 1 Introduction

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

11 Chapter 2 Literature Review

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.

12 Chapter 2 Literature Review

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

13 Chapter 2 Literature Review 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

14 Chapter 2 Literature Review 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

15 Chapter 2 Literature Review 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.

16 Chapter 2 Literature Review

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

17 Chapter 2 Literature Review 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).

18 Chapter 2 Literature Review

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

19 Chapter 2 Literature Review 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

20 Chapter 2 Literature Review 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).

21 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

22 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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

23 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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).

24 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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 Continued …….

25 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

Figure 3.1 Continued …….

26 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

27 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

Table 3.1 Description of PRECIS RCM experiment

Experiment Driving Fields Resolution Number Time Period (km) of Grids

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

(T  T) 2 STDV  i i (3.1) N 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 ( STDVP ) is normalized by the average precipitation ( P ) because STDVP is affected by the mean, so that the CV is more independent measure of interannual variability.

STDV CV  P (3.2) P

28 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

29 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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).

30 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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 Continued …….

31 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

Figure 3.4 Continued …….

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

32 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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.

33 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

Continued …….

34 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

Figure 3.5 Continued …….

35 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

36 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

30 30 25 25 20 20 15 15 10 5 10

Temperature (˚C) 0 Temperature (˚C) 5 -5 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Region 1 Month Region 4 Month

CRU PRECIS-Had HadAM3P CRU PRECIS-Had HadAM3P

40 10 35 5 30 0 25 -5 20 -10 15 10 -15 Temperature (˚C) 5 Temperature-20 (˚C) 0 -25 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 2 Region 5

CRU PRECIS-Had HadAM3P CRU PRECIS-Had HadAM3P

20 25 15 10 20 5 15 0 -5 10 -10 5 Temperature (˚C) -15 Temperature (˚C) -20 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Region 3 Month Month Region 6 CRU PRECIS-Had HadAM3P CRU PRECIS-Had HadAM3P

40 35 30 25 20 15 10 Temperature (˚C) 5 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Region 7 CRU PRECIS-Had HadAM3P

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 Chapter 3 Analyses ofP:RECIS RCM Climate Change Scenarios

-, ...... ------. 40 ...... --...... --...... --...... --.--..--. t ~~r ;; ~ 20 ~ 30 ~ 15 20 : 10 i" ~ 5 J 10 0 ...... 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region1 Region4

! 4--CRU -. .. -. .PRECIS-ERAI' I 4--CRU...... PRECIS-ER~I

""'''''''''''''''''''''''''---'''''---'' ---...... 10,...... --....--....----..--...... f" Q' 40 , . ;> 5 ~ 0, ~ 30 ;; -5 20 : I ~ -10 ~-15 .:-20 11: 1 -25 1 : Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec I Month Month Region2 Region5

I 4--CRU . - .PRECIS-ERAI 1 4-- CRU . . .. - . .PRECIS-ERAJ

~ .-.

I 20

."""'---"'''''' Q ;> ~: ---''''''~'~':'' :;,. 10 ~ ... ," '. ~ 15 . ~ 0 . . "~ 10 " .101 .'. ~~ J 5 -20 0I Jan Feb Mar Apr May Jun Jut Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Moolth Region3 Region6

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

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

J 1: l' : Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Region7

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

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

38 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

1.50 1.50

1.00 1.00

0.50 0.50 STDV STDV (˚C) STDV (˚C)

0.00 0.00 R1 R2 R3 R4 R5 R6 R7 R1 R2 R3 R4 R5 R6 R7 Subregions Subregions JFM JAS CRU PRECIS HadAM3P CRU PRECIS HadAM3P

1.50 1.50

1.00 1.00

0.50 0.50 STDV (˚C) STDV STDV (˚C)

0.00 0.00 R1 R2 R3 R4 R5 R6 R7 R1 R2 R3 R4 R5 R6 R7 Subregions Subregions AMJ OND CRU PRECIS HadAM3P CRU PRECIS HadAM3P

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) Region PRECIS-Had HadAM3P PRECIS-ERA 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

39 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

40 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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.

41 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

Continued …….

42 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

Figure 3.10 Continued …….

43 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

44 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

2.5 10

2 8

1.5 6

1 4

0.5 2

Precipitation (mm/day)0 Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month Region 4 Month Region 1 CRU PRECIS-Had HadAM3P CRU PRECIS-Had HadAM3P

5 6

4 5 4 3 3 2 2 1 1

Precipitation0 (mm/day) Precipitation0 (mm/day) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 2 Region 5 CRU PRECIS-Had HadAM3P CRU PRECIS-Had HadAM3P

5 12 4 10 8 3 6 2 4 1 2 Precipitation (mm/day) 0 Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 3 Region 6 CRU PRECIS-Had HadAM3P CRU PRECIS-Had HadAM3P

14 12 10 8 6 4 2

Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Region 7 CRU PRECIS-Had HadAM3P

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 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

2.5 8 7 2 6 1.5 5 4 1 3 0.5 2 1 Precipitation (mm/day) 0 Precipitation0 (mm/day) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month Region 4 Month Region 1 CRU PRECIS-ERA CRU PRECIS-ERA

5 5

4 4

3 3

2 2

1 1 Precipitation (mm/day) 0 Precipitation0 (mm/day) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 2 Region 5

CRU PRECIS-ERA CRU PRECIS-ERA

5 10

4 8

3 6

2 4

1 2 Precipitation (mm/day) 0 Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 3 Region 6 CRU PRECIS-ERA CRU PRECIS-ERA

Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Region 7 CRU PRECIS-ERA

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

46 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

1 00 '"''

" O.

~PRECIS .II>.HadAM3P I 10 CRU . PRECIS.II>. H;;dAM 3P J

--,

100 1" ,"":"'"'''''';'''''''''''':''''''''''';'''''''''''':''''''''''~'''''''''''

# 0 751-_n _~_nnn:n_n_n:nnnn:---nn" f : ~ 0 ;11:' . . I. ::: [; 025. n_n_;nn: - -: On. n, , . ~--i--:---i---~---:---:-_.:-_I- ; ~---: , , ' 0.00 l -- --; 0.00: l~iiIoI~:.:m~...~'---u_+ m~ +.._u '-"._"", ,,; I R1 R2 R3 R4 R5 R6 R7 Rl R2 R3 R4 R5 R6 R7

Subregions Subregions ON AMJ FCRU .PRECIS .II>.HadAM3P I 10CRU . PRECIS .II>.HadA~

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 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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.

1.00

0.80

0.60

0.40

Wetday frequency 0.20

0.00 R1 R2 R3 R4 R5 R6 R7 Region

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

48 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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.

49 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

50 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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

51 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

40 40

30 30

20 20

10 10 0

Temperature (˚C) 0 Temperature (˚C) -10 -10 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Region 1 Month Region 4 Month

Present Temperature Future Temperature Present Temperature Future Temperature

40 20

30 10

20 0

10 -10

0 -20 Temperature (˚C) Temperature (˚C) -10 -30 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 2 Region 5

Present Temperature Future Temperature Present Temperature Future Temperature

20 40

10 30

0 20

-10 10 -20 Temperature (˚C) 0 Temperature (˚C) -30 -10 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Region 3 Month Month Region 6 Present Temperature Future Temperature Present Temperature Future Temperature

10 Temperature (˚C) 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Region 7 Present Temperature Future Temperature

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 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

2 10

6 1 4

Precipitation0 (mm/day) Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Region 1 Month Region 4 Month

Present Precipitation Future Precipitation Present Precipitation Future Precipitation

5 8

4 6 3 4 2 2 1 Precipitation (mm/day) Precipitation0 (mm/day) 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 2 Region 5 Present Precipitation Future Precipitation Present Precipitation Future Precipitation

5 10

4 8

3 6

2 4

1 2 Precipitation (mm/day) Precipitation0 (mm/day) 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month Region 3 Region 6 Present Precipitation Future Precipitation Present Precipitation Future Precipitation

16 14 12 10 8 6 4 2 Precipitation (mm/day)0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Region 7 Present Precipitation Future Precipitation

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 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios

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)

Region Temperature Change (˚C) Precipitation Change (%) 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

54 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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

55 Chapter 3 Analyses of PRECIS RCM Climate Change Scenarios 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 %.

56 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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.

57 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

Table 4.1 Characteristics of study area

Parameters River Basins 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

58 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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.

59 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

Temperature (°C) Precipitation (mm)

Station Date Max. Min. Mean Hr.12 of Hr.0 To Past 24 Hrs. Hr.3 To previous Hr.3 of Preceding Hr.12 Event date to Hr. Date Hr. 3 Of of Date 0 of date 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 Astore 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 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 A Gilgit 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 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 Skardu 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 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 Astore 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 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 B Gilgit 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 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 Skardu 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 30 25.3 13 19.15 0.0 0.0 0.0 0.0 A: Event of October, 1987, B: Event of August 1997

60 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

a) Discharge during October 1-20, 1987

600

400 /s) 3

Discharge(m 200

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Day

Hunza Gilgit Astore

b) Discharge during August 20 to September 8, 1997 1200

800 /s) 3

400 Discharge (m Discharge

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Day

Hunza Gilgit Astore

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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.

62 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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) River Basin PRECIS-Had PRECIS-ERA 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 (%) River Basin PRECIS-Had PRECIS-ERA Winter Summer Winter Summer Astore 135 57 361 133

Gilgit 43 69 103 44

Hunza 167 217 297 210

63 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

15 a iHunza RlVerBas;n ]CI

/' --::-'>(- ' . -:-'X:.~ "- ""', 0 "' ,,:,.x., '''''XIo, ::; -5 - -.x' .... ;;; ,- -"" ''> :;; -10 /,~ ~ E ;:../" A ~ -15 ,,;,; "" - ' ' ,x,- -20 " x ' .-- '" ,; ~'; , , ,, X - " /' '

.30 IAN FEB MAR APR MAY JlIN IUL AUG SEP OCT NOV DEC Mdnth

1- - - - PRECIS.fud- CRU. . . X . . . PRECIS-ERA I

15 b) Gilglt River Bas in 10

/- -::: :x =-. :--.""", '" ,x' "- '" " ' i:: 0 ,~ ,-",~,., ~, ~ -5 '" "'..,' '" :;; ,;.x' .'). CL '-' E -10 ~, f- .T" ~ -15 ,'/ " .,' -,x;,/ '. , X '" ~ "- -20 ",'" <:,'. ,; ," "x, -25 IAN FEB MAR APR MAY IUN IUL AUG SEP OCT NOV DEC Month

1- - - - PRECIS-fud50~ CRU. . . X - - -PRECIS.ERAI

20 c) As tore R;yerBasll1 15 :--:_--::-.~, 10 "',' " '" E ~/.,; .,,X' . . 'X" .,- ~, " '" ::; '" ,x' ',' - ;;; 0 ,; . '~ :;; ,- " x, "- CL -(.,;,~' , '" ~.E -5 f- ..:-~. ~'" -10 ..-x,-,'/ , , "x x" / / , / . -15 '"

-20 IAN FEB MAR APR' MAY IUN WL AUG SEP OCT NOV DEC Month

1- - - -PREeIS-fud50-eRU . "X' - 'PRECIS-ERAJ

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

64 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

7rI a)Hullza RLverB,HIIL 6 1 ,x.. ."=., ! '- "'"- ~-" E - ',-" ..§. 4 /' x. .:: , 0 'x '., >< ' '." 'j 3 '" '", :9- x '" ~ x- - . . . x' .~.e. -, / ", cL .....-- 1 '" "'X. ------..... '" "' ""'><'-"'-:'><""'X- -- / 0 , , ; TAN FEB MAR APR MAY TUN TUL AUG SEP OCT NOV DEC Month

1- - - -PREClS-H,d-CRD"'X" 'PRECIS-ERA1

...... -..-..--...... -...... -.._---..--...... 6 r ! b) Gilglt RlverBas 111 5 -; x =. .- -, ~ 4 , r -.. " : '" " , .S ' '" '" .Q 3 ,'", 'x." '" ,>y "" \ .' '" 'x, .x lc .' '" \, Il.. x'", ~ "'. '" I,./"" \ x, "

0-1--- r ; TAN FEB MAR APR MAY TUN TUL AUG SEP OCT NOV DEC Month

1- - - - PREClS-.Hod"- CRD. - . X . . . PRECIS-ERA 1

12 ,"""""" -...... --......

c)Astore RIVerBas., 10 .x '"= . ,,,' ~ 8 S " ;x x x :i (, 0 .§- 4 /< . . ::: Il.. /"' ""' , x.,.. "'~-_. / ',"''''..x' 'X',.' / 2 I '- - - -, ". .' / '" ---- , ~ 'x '-- -- 0 -r--i TAN FEB MAR APR MAY TUN TUL AUG SEP OCT NOV DEC Month

1- - - -PREcrS.H,dSO-CRD"'X" -PREClS.ERA1

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

66 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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.

RF  pcorr.rfcf .P if T  TT (4.1)

SF  pcorr.sfcf .P if T  TT (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

67 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling 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).

Sm  cfmaxT  TT  (4.3)

Rmw  cfr.cf max .(TT  T) (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).

Gm  gmeltT  TT  (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  SM  R  IN   (4.6)  FC 

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).

SM E   E if SM  LP  FC (4.7) a LP  FC p

Ea  E p if SM  LP  FC (4.8)

68 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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).

 FC  SM  C f  CFLUX    (4.9)  FC 

1ALFA Q0  K f UZ (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).

Q1  K s  LZ (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

69 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

NS Y  (4.12) 1 RE

where

iN 2 Qs i Qo i NS  1 i1 (4.13) iN 2 Qo i- Qo  i1

iN  Qs i  Qo i RE 100 i1 iN (4.14)  Qo i i1 where i is the time step, N is the total number of time steps, Qs represents simulated discharge, Qo is observed discharge and Qo 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.

70 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

72 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling 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

73 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling 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.

74 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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)

d 0.09 1.0 0 -0.5 4 3.0 0.5 SMHI , (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)

Killingtveit and 75-300 0.7-1 -1-2 -1-2 3-6 0.5-1 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 Parameter River basins 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

75 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

Met

- 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 HBV 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

ERA

- 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 HBV 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 PRECIS -

HBV 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

76 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

(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

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

77 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

GMELT=3.5 GMELT=3.5 1 1 1 DTTM=0 DTTM=1 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 0 TT 0 TT 0 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 -1 -1 -1 900 1200 1500 0.5-0.55 900 1200 1500 0.5-0.55 900 1200 1500 0.5-0.55 FC FC FC

GMELT=4 GMELT=4 GMELT=4 1 1 1 DTTM=-1 DTTM=0 DTTM=1 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 0 TT 0 TT 0 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 -1 -1 -1 0.5-0.55 900 1200 1500 0.5-0.55 900 1200 1500 0.5-0.55 900 1200 1500 FC FC FC

GMELT=4.5 GMELT=4.5 GMELT=4.5 1 1 1 DTTM=-1 DTTM=0 DTTM=1 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 0 TT 0 TT 0 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 -1 -1 -1 900 1200 1500 0.5-0.55 900 1200 1500 0.5-0.55 900 1200 1500 0.5-0.55 FC FC FC

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

GMELT=3.5 GMELT=3.5 -0.8 -0.8 -0.8 DTTM=0 DTTM=1 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 -0.3 TT -0.3 TT -0.3 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 0.2 0.2 0.2 100 600 1100 0.5-0.55 100 600 1100 0.5-0.55 100 600 1100 0.5-0.55 FC FC FC

GMELT=4 GMELT=4 GMELT=4 -0.8 -0.8 -0.8 DTTM=-1 DTTM=0 DTTM=1 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 -0.3 TT -0.3 TT -0.3 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 0.2 0.2 0.2 0.5-0.55 100 600 1100 0.5-0.55 100 600 1100 0.5-0.55 100 600 1100 FC FC FC

GMELT=4.5 GMELT=4.5 GMELT=4.5 -0.8 -0.8 -0.8 DTTM=-1 DTTM=0 DTTM=1 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 -0.3 TT -0.3 TT -0.3 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 0.2 0.2 0.2 100 600 1100 0.5-0.55 100 600 1100 0.5-0.55 100 600 1100 0.5-0.55 FC FC FC

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

GMELT=3.5 GMELT=3.5 0.9 0.9 0.9 DTTM=-1.5 DTTM=-2 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 0.4 TT 0.4 TT 0.4 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 -0.1 -0.1 -0.1 1100 1300 1500 0.5-0.55 1100 1300 1500 0.5-0.55 1100 1300 1500 0.5-0.55 FC FC FC

GMELT=4 GMELT=4 GMELT=4 0.9 0.9 0.9 DTTM=-1 DTTM=-1.5 DTTM=-2 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 0.4 TT 0.4 TT 0.4 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 -0.1 -0.1 -0.1 0.5-0.55 1100 1300 1500 0.5-0.55 1100 1300 1500 0.5-0.55 1100 1300 1500 FC FC FC

GMELT=4.5 GMELT=4.5 GMELT=4.5 0.9 0.9 0.9 DTTM=-1 DTTM=-1.5 DTTM=-2 0.85-0.9 0.85-0.9 0.85-0.9 0.8-0.85 0.8-0.85 0.8-0.85 0.75-0.8 0.75-0.8 0.75-0.8 0.4 TT 0.4 TT 0.4 TT 0.7-0.75 0.7-0.75 0.7-0.75 0.65-0.7 0.65-0.7 0.65-0.7 0.6-0.65 0.6-0.65 0.6-0.65 0.55-0.6 0.55-0.6 0.55-0.6 -0.1 -0.1 -0.1 1100 1300 1500 0.5-0.55 1100 1300 1500 0.5-0.55 1100 1300 1500 0.5-0.55 FC FC FC

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

a) HBV-Met 2000

1500 /s) 3

1000

500 Discharge Discharge (m

0 1981 1982 1983 1984 1985 1986 1987 1988 1989

Simulated discharge (Qs) Observed discharge (Qo)

b) HBV-ERA 2000

1500 /s) 3

1000

500 Discharge Discharge (m

0 1981 1982 1983 1984 1985 1986 1987 1988 1989

Simulated discharge (Qs) Observed discharge (Qo)

c) HBV-PRECIS 2000

1500 /s) 3

1000

500 Discharge Discharge (m

0 1981 1982 1983 1984 1985 1986 1987 1988 1989

Simulated discharge (Qs) Observed discharge (Qo)

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

81 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

a) HBV-Met :s 11200 1000 ~.," ~

"'".~ 800 ]: 600 -,.," .~v, 400 ~ {; 200 ~ a '"'-' .,: a 200 400 600 800 1000 1200

Accum ulated 0 bserved discharge (1J3m3/s)

b) HBV-ERA OJ, ~ 1200 ~ 7. 1000 ...,2" "* 800 '6 ~ 600

]'.. 400 "'" ~ 200 :5 E i'i a Q -0: a 200 400 600 800 1000 1200

Accumulated observed d~scharge (1J3m3/s)

c) HBV-PRECIS OJ, ~ 1200 S2

E' 1000 ;7: 800 '6 ':i: 600 ~ :5 'ViE 400

~..., 200 :5 E i'i a .:;. -0: a 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 basin during calibration period

82 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

1200 a) Hunza river basin 1000 /s) 3 800 600 400

Discharge (m Discharge 200 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

HBV-PRECIS Observed HBV-Met HBV-ERA

1000 b) Gilgit river basin 800 /s) 3 600

400

Discharge (m Discharge 200

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

HBV-PRECIS Observed HBV-Met HBV-ERA

500 c) Astore river basin 400 /s) 3 300

200

Discharge (m Discharge 100

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

HBV-PRECIS Observed HBV-Met HBV-ERA

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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)

Y  Y ARD  100 a c (4.15) Yc 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

84 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

85 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling 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 PRECIS-ERA PRECIS-Had Observations

HBV-Met 0.87 0.65(25) 0.53(39)

Hunza HBV-ERA 0.49(45) 0.89 0.73(18)

HBV-Had 0.56(27) 0.86(12) 0.77

HBV-Met 0.82 0.55(33) 0.62(24)

Gilgit HBV-ERA 0.57(24) 0.75 0.63(16)

HBV-Had 0.67(9) 0.64(14) 0.74

HBV-Met 0.67 0.20(70) 0.55(18)

Astore HBV-ERA 0.31(46) 0.58 0.37(36)

HBV-Had 0.57(6) 0.35(43) 0.61

86 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

Return period (years) 1 2 5 10 20 50 3,000 HBV-Met a) Hunza River Basin 2,500 HBV-PRECIS /s)

3 2,000 Observed 1,500

1,000 Discharge (m 500

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

Return period (years) 5 1 2 10 20 50 3,000 HBV-Met 2,500 b) Gilgit River Basin HBV-PRECIS /s) 3 2,000 Observed 1,500 1,000

Discharge (m 500

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

Return period (years)

1 2 5 10 20 50 2000 HBV-Met c) Astore River Basin 1500 HBV-PRECIS /s) 3 Observed 1000

500 Discharge (m

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate 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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

2000 2000 a) HBV-Met 1500 1500

1000 1000

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

2000 2000 b) HBV-ERA 1500 1500

1000 1000

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

2000 2000 c) HBV-PRECIS 1500 1500

1000 1000

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

2000 2000 a) HBV-Met 1500 1500

1000 1000

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

2000 2000 b) HBV-ERA 1500 1500

1000 1000

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

2000 2000 c) HBV-PRECIS

1500 1500

1000 1000

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

1000 1000 a) HBV-Met

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361

1000 1000 b) HBV-ERA

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

1000 1000 c) HBV-PRECIS

500 500

0 0 1 31 61 91 121 151 181 211 241 271 301 331 361 Day

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 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling

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

91 Chapter 4 PRECIS Simulations as Input to Hydrological Modelling 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.

92 Chapter 5 Climate Change Impact on Water Resources

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

93 Chapter 5 Climate Change Impact on Water Resources 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)

River Temperature Change (˚C) Precipitation Change (%) 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

94 Chapter 5 Climate Change Impact on Water Resources

10 5 a) Hunza River Basin 0 -5 -10 -15 -20 -25 Mean temperature (˚C) temperature Mean -30 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Present Temperature Future Temperature

20 b) Gilgit River Basin 10

-10

-20

Mean temperature (˚C) temperature Mean -30 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Present Temperature Future Temperature

20 15 c) Astore River Basin 10 5 0 -5 -10 -15 Mean temperature (˚C) temperature Mean -20 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Present Temperature Future Temperature

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 Chapter 5 Climate Change Impact on Water Resources

10 a) Hunza River Basin 8

Precipitation (mm/day) 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Present Precipitation Future Precipitation

10 b) Gilgit River Basin 8

Precipitation (mm/day) 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Present Precipitation Future Precipitation

12 10 c) Astore River Basin 8 6 4 2

Precipitation (mm/day) 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Present Precipitation Future Precipitation

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 Chapter 5 Climate Change Impact on Water Resources

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 (T f ,daily ) and daily precipitation ( Pf ,daily ) time series are constructed by using equations 5.1 and 5.2, respectively.

T f ,daily  To,daily  T f ,monthly  Tp,monthly  (5.1)

Pf ,monthly Pf ,daily  Po,daily (5.2) Pp,monthly

Where To,daily is the observed daily temperature, Po,daily is the observed daily precipitation,

T f ,monthly is the mean monthly PRECIS simulated future temperature, Tp,monthly is the mean monthly PRECIS simulated present temperature, Pf ,monthly is the mean monthly PRECIS

97 Chapter 5 Climate Change Impact on Water Resources

simulated future precipitation and Pp,monthly 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

98 Chapter 5 Climate Change Impact on Water Resources 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.

99 Chapter 5 Climate Change Impact on Water Resources

1500 a) Hunza river basin /s) 3 1000

500 Discharge Discharge (m 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Future Discharge-100% glaciation Future Discharge-50% glaciation Future Discharge-0% glaciation Present simulated discharge

1500 b) Gilgit river basin /s) 3 1000

500 Discharge Discharge (m 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Future Discharge-100% glaciation Future Discharge-50% glaciation Future Discharge-0% glaciation Present simulated discharge

500 c) Astore river basin

/s) 400 3 300 200 100 Discharge Discharge (m 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Future Discharge-100% glaciation Future Discharge-50% glaciation Future Discharge-0% glaciation Present simulated discharge

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 Chapter 5 Climate Change Impact on Water Resources

2000 a) Hunza river basin /s) 3 1500

1000

500 Discharge Discharge (m 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Future Discharge-100% glaciation Future Discharge-50% glaciation Future Discharge-0% glaciation Present simulated discharge

1500 b) Gilgit river basin /s) 3 1000

500 Discharge Discharge (m 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Future Discharge-100% glaciation Future Discharge-50% glaciation Future Discharge-0% glaciation Present simulated discharge

500 c) Astore river basin

/s) 400 3 300 200 100 Discharge Discharge (m 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month

Future Discharge-100% glaciation Future Discharge-50% glaciation Future Discharge-0% glaciation Present simulated discharge

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 Chapter 5 Climate Change Impact on Water Resources

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 Under 50 % Without glaciers day glaciations reduction in glaciers

Hunza 26 -27 -81 Met - Gilgit 44 -30 -96 HBV

Astore 17 -27 -72

Hunza 65 -13 -91 Gilgit 40 -20 -78 PRECIS - Astore 46 -24 -93 HBV

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

102 Chapter 5 Climate Change Impact on Water Resources 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.

103 Chapter 5 Climate Change Impact on Water Resources

Return period (years)

1 2 5 10 20 50 3500 Future-100% glaciation Future-50% glaciation 3000 a) Hunza River Basin Future-0% glaciation Present simulated

/s) 2500 3 2000 1500 1000 Discharge(m 500 0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

Return period (years)

1 2 5 10 20 50 2500

Future-100% glaciation Future-50% glaciation b) Gilgit River Basin 2000

/s) Future-0% glaciation Present simulated 3 1500

1000

Discharge(m 500

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

Return period (years)

1 2 5 10 20 50 1500 Future-100% glaciation Future-50% glaciation c) Astore River Basin Future-0% glaciation Present simulated /s) 3 1000

500 Discharge(m

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

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 Chapter 5 Climate Change Impact on Water Resources

Return period (years)

1 2 5 10 20 50 3500 Future-100% glaciation Future-50% glaciation 3000 a) Hunza River Basin Future-0% glaciation Present simulated

/s) 2500 3 2000 1500 1000 Discharge(m 500 0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

Return period (years)

1 2 5 10 20 50 2500 Future-100% glaciation Future-50% glaciation b) Gilgit River Basin 2000 Future-0% glaciation Present simulated /s) 3 1500

1000

Discharge(m 500

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

Return period (years)

1 2 5 10 20 50 1500 Future-100% glaciation Future-50% glaciation c) Astore River Basin Future-0% glaciation Present simulated /s) 3 1000

500 Discharge(m

0 -2 -1 0 1 2 3 4 Reduced Gumbel variate

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 Chapter 5 Climate Change Impact on Water Resources

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) Model River Basin Glaciation Scenario 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 PRECIS - Astore 521.6 278.6 43.0 20.3 30.4 28.5 HBV

1573.6 866.5 252.4 256.6 298.6 346.5 Hunza (1868.7) (1182.3) (500.4) (1169.5) (1256.7) (1016.8) Met - Gilgit 1282.7 605.1 34.6 144.3 83.1 40.3 HBV

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

106 Chapter 5 Climate Change Impact on Water Resources 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.

107 Chapter 6 Conclusions and Recommendations for Future Work

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

108 Chapter 6 Conclusions and Recommendations for Future Work

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

109 Chapter 6 Conclusions and Recommendations for Future Work

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.

110 Chapter 6 Conclusions and Recommendations for Future Work

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.

111 References

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.

112 References

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.,

113 References

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

114 References

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.

115 References

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

116 References

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 socioeconomic 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.

117 References

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.

118 References

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.

119 References

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.

120 References

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.

121 References

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.

122 References

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.

123 References

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.

124 References

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.

125 References

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.

126 References

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.

127 References

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.

128 References

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.

129 References

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.

130 References

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.

131

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.

132