Water availability, demand and adaptation option assessment of the Ganges Basin under climate change Technical Report – Hydrological Modelling Koshi, Tamkoshi, Dudhkoshi, Tamor Basins and Ganga Basin

Department of Civil Engineering Indian Institute Of Technology, Delhi Hauz Khas, New Delhi 110016, India

Water availability, demand and adaptation option assessment of the Ganges Basin under climate change

Technical Report – Hydrological Modelling Koshi, Tamkoshi, Dudhkoshi, Tamor Basins and Ganga Basin

Contract no: Project Funded by: November 2013

Department of Civil Engineering Indian Institute Of Technology, Delhi Hauz Khas, New Delhi 110016, India

Table of Contents TABLE OF CONTENTS ...... I LIST OF FIGURES ...... III LIST OF TABLES ...... V EXECUTIVE SUMMARY ...... VI

CHAPTER 1 - INTRODUCTION ...... 8

THE KOSHI RIVER BASIN PROFILE ...... 8 PHYSIOGRAPHY ...... 8 CLIMATE ...... 9 SOILS ...... 10 LANDUSE ...... 10 MAJOR TRIBUTARIES ...... 10 IRRIGATION ...... 10

CHAPTER 2 - SWAT MODEL SETUP FOR THE KOSHI BASIN AND CURRENT BASELINE SIMULATION . 12

METHODOLOGY ...... 12 SOIL AND WATER ASSESSMENT TOOL (SWAT) MODEL ...... 12 Advantages of the SWAT model ...... 13 DEVELOPMENT OF HYDROLOGIC MODEL FOR THE KOSHI RIVER BASIN ...... 15 Model Performance ...... 16 MAPPING THE KOSHI RIVER BASIN ...... 18 Basin Demarcation ...... 18 Watershed (sub-basin) Delineation ...... 19 Land Cover/Land Use Layer ...... 20 Soil Layer ...... 21 Hydro-Meteorological and Water resources structures data ...... 22 SWAT MODEL PERFORMANCE FOR THE KOSHI BASIN ...... 24 THE HYDROLOGIC SIMULATION WITH OBSERVED WEATHER ...... 27 Temporal distribution of major water balance components...... 28 Flow Duration Curve ...... 30

CHAPTER 3 - CLIMATE CHANGE SCENARIOS USING BCCR IPCC AR5 RCP WEATHER DATA ...... 36

CLIMATE CHANGE DATA USED FOR HYDROLOGICAL MODELLING ...... 36 THE HYDROLOGIC SIMULATION WITH CLIMATE CHANGE SCENARIOS ...... 37 Seasonal Change in water balance components for RCP 4.5 and RCP 8.5 Scenarios ...... 42 IMPACT ASSESSMENT ...... 44 DROUGHT ANALYSIS ...... 44 FLOOD ANALYSIS ...... 46 Average peak discharge at 99th percentile...... 47

i

Flow Duration Curve and Impact on Reservoirs ...... 47 LIMITATIONS OF THE STUDY ...... 53 ASSUMPTIONS ...... 53 Data Limitations ...... 53 Model limitations ...... 54

CHAPTER 4 - CLIMATE CHANGE SCENARIOS USING - DELTA CHANGE METHOD WEATHER DATA ... 56

CLIMATE CHANGE DATA USED FOR HYDROLOGICAL MODELLING ...... 56 THE HYDROLOGIC SIMULATION WITH CLIMATE CHANGE SCENARIOS ...... 56 Flow Duration Curve and Impact on dependable flow for Koshi Basin ...... 56 Climate Change Impact on flow for Ganga Basin...... 66

CHAPTER 5 - FUTURE DEMAND SCENARIOS AND IMPACT OF CLIMATE CHANGE ...... 82

FUTURE WATER DEMAND SCENARIO ...... 82 THE HYDROLOGIC SIMULATION WITH CLIMATE CHANGE SCENARIOS ...... 82 Flow Duration Curve and Impact on dependable flow for Koshi Basin ...... 83

CHAPTER 5 - SUMMARY AND CONCLUSIONS ...... 93

ii

List of Figures FIGURE 1 : GEOGRAPHICAL CONTEXT OF THE KOSHI RIVER BASIN ...... 9 FIGURE 2 : STUDY AREA - THE KOSHI BASIN ...... 15 FIGURE 3 : DIGITAL ELEVATION MODEL OF THE KOSHI RIVER BASIN ...... 18 FIGURE 4 : BASIN DELINEATION USING DEM FOR THE KOSHI RIVER BASIN ...... 19 FIGURE 5 : SUB BASIN DELINEATION USING DEM FOR THE KOSHI RIVER BASIN ...... 19 FIGURE 6 : LANDUSE MAP FOR THE KOSHI RIVER BASIN...... 20 FIGURE 7 : SOIL MAP FOR THE KOSHI RIVER BASIN ...... 21 FIGURE 8 : : METEOROLOGICAL LOCATIONS FOR THE KOSHI RIVER BASIN ...... 22 FIGURE 9 : STREAM FLOW MEASUREMENT LOCATIONS MAP OF KOSHI RIVER BASIN ...... 23 FIGURE 10 : SWAT OUTPUT COMPARISON LOCATIONS AND MODEL EFFICIENCY PARAMETERS FOR THE KOSHI BASIN ...... 26 FIGURE 11 : SIMULATED AVERAGE WATER BALANCE COMPONENTS FOR KOSHI RIVER BASIN ...... 28 FIGURE 12 : SIMULATED AVERAGE MONTHLY WATER BALANCE COMPONENTS FOR KOSHI RIVER BASIN ...... 29 FIGURE 13 : SIMULATED AVERAGE ANNUAL WATER BALANCE COMPONENTS FOR KOSHI RIVER BASIN ...... 30 FIGURE 14 : FLOW DURATION CURVES AT VARIOUS LOCATIONS ON KOSHI RIVER BASIN ...... 31 FIGURE 15 :CLIMATE CHANGE GRID LOCATIONS IN THE KOSHI RIVER BASIN ...... 37 FIGURE 16 ANNUAL WATER BALANCE COMPONENTS FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5) FOR THE KOSHI RIVER BASIN ...... 38 FIGURE 17 ANNUAL WATER BALANCE COMPONENTS FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 39 FIGURE 18 MONTHLY WATER BALANCE COMPONENTS FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 40 FIGURE 19 SPATIAL DISTRIBUTION OF CHANGE IN WATER BALANCE COMPONENTS TOWARDS MID CENTURY (IPCC AR5 RCP 4.5 AND 8.5) FOR THE KOSHI RIVER BASIN ...... 41 FIGURE 20 SEASONAL CHANGE IN WATER BALANCE (MM) EXPRESSED AS PERCENT CHANGE IN PRECIPITATION (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 43 FIGURE 21 CHANGE IN SPATIAL DISTRIBUTION OF SOIL MOISTURE DEFICIT WEEKS FOR BL AND MC SCENARIOS (IPCC AR5 RCP 4.5) FOR THE KOSHI RIVER BASIN – SOUTH WEST MONSOON (JJAS) ...... 45 FIGURE 22 CHANGE IN SPATIAL DISTRIBUTION OF SOIL MOISTURE DEFICIT WEEKS FOR BL AND MC SCENARIOS (IPCC AR5 RCP 8.5) FOR THE KOSHI RIVER BASIN – SOUTH WEST MONSOON (JJAS) ...... 46 FIGURE 23 SPATIAL VARIATION IN CHANGE IN STREAM DISCHARGE AT 99TH PERCENTILE FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 47 FIGURE 24 LOCATIONS WHERE DEPENDABLE FLOW IS ASSESSED FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 48 FIGURE 25 DEPENDABLE FLOW IS ASSESSED FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR SEVEN LOCATIONS IN KOSHI RIVER BASIN ...... 49 FIGURE 26 LOCATIONS WHERE DEPENDABLE FLOW IS ASSESSED FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 57 FIGURE 27 SIMULATED DEPENDABLE FLOW FOR OBSERVED AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) FOR SEVEN LOCATIONS IN KOSHI RIVER BASIN ...... 58 FIGURE 28 SIMULATED 75% DEPENDABLE FLOW FOR OBSERVED AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) FOR SEVEN LOCATIONS IN KOSHI RIVER BASIN ...... 62 FIGURE 29 LOCATIONS WHERE DEPENDABLE FLOW IS ASSESSED FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) FOR THE GANGA RIVER BASIN ...... 66 FIGURE 30 SIMULATED DEPENDABLE FLOW (BASED ON AVERAGE MONTHLY FLOWS) FOR OBSERVED CLIMATE AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) AT VARIOUS LOCATIONS IN GANGA RIVER BASIN ...... 67 iii

FIGURE 31 SIMULATED 75% DEPENDABLE FLOW FOR OBSERVED AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) FOR VARIOUS LOCATIONS IN GANGA RIVER BASIN ...... 77 FIGURE 32 LOCATIONS OF PLANNED RESERVOIRS AND LOCATIONS WHERE WHERE DEPENDABLE FLOW IS ASSESSED FOR BL AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5) FOR THE KOSHI RIVER BASIN ...... 83 FIGURE 33 SIMULATED DEPENDABLE FLOW FOR OBSERVED AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) AT MULGHAT IN KOSHI RIVER BASIN ...... 84 FIGURE 34 SIMULATED DEPENDABLE FLOW FOR OBSERVED AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) AT CHARATA IN KOSHI RIVER BASIN ...... 85 FIGURE 35 SIMULATED 75% DEPENDABLE FLOW FOR OBSERVED BAU, OBSERVED AND MC CLIMATE SCENARIOS (IPCC AR5 RCP 4.5 AND RCP 8.5, STATISTICAL DOWNSCALING) FOR MULGHAT AND CHATARA IN KOSHI RIVER BASIN WITH 3 DEMAND SCENARIOS ...... 88

iv

List of Tables TABLE 1 ELEVATION SUMMARY –STUDY AREA ...... 18 TABLE 2 LANDUSE CATEGORIES –OF THE KOSHI RIVER BASIN ...... 20 TABLE 3 SOIL TEXTURE OF THE KOSHI RIVER BASIN ...... 21 TABLE 4 STREAM FLOW MEASUREMENT STATIONS –KOSHI RIVER BASIN ...... 22 TABLE 5 SWAT OUTPUT COMPARISON LOCATIONS AND MODEL EFFICIENCY PARAMETERS ...... 25 TABLE 6 FLOW DEPENDABILITY – SUB-BASINS OF KOSHI RIVER BASIN ...... 30 TABLE 7: OVERVIEW OF REPRESENTATIVE CONCENTRATION PATHWAYS (RCPS) ADOPTED BY IPCC AR5 ...... 36 TABLE 8 SUMMARY OF SEASONAL CHANGE IN WATER BALANCE (MM) EXPRESSED AS PERCENT CHANGE IN PRECIPITATION FOR THE STUDY AREA (BASELINE TO MID CENTURY) – RCP 4.5 AND RCP 8.5 ...... 42 TABLE 9 DEPENDABLE FLOW AT VARIOUS LOCATIONS (MCM) FOR IPCC AR5 RCP 4,5 AND RCP 8.5 SCENARIOS (BASELINE, AND MID CENTURY) ...... 52 TABLE 10 MODEL SELECTION BASED ON 10TH AND 90TH PERCENTILE VALUES OF PROJECTED CHANGES IN P AND T FROM 1961-1990 TO 2021-2050...... 56 TABLE 11 DEPENDABLE FLOW AT VARIOUS LOCATIONS (CMS) FOR IPCC AR5 RCP 4,5 AND RCP 8.5 SCENARIOS (BASELINE, AND MID CENTURY) ...... 61 TABLE 12 DEPENDABLE FLOW AT VARIOUS LOCATIONS ON GANGA (CMS) FOR IPCC AR5 RCP 4,5 AND RCP 8.5 SCENARIOS (BASELINE, AND MID CENTURY) ...... 74 TABLE 13 DETAILS OF THE FUTURE DEVELOPMENT PROJECTS ...... 82 TABLE 14 SCENARIOS USED IN THE STUDY ...... 82 TABLE 15 DEPENDABLE FLOW AT MULGHAT AND CHATARA LOCATIONS (CMS) FOR IPCC AR5 RCP 4,5 AND RCP 8.5 SCENARIOS (BASELINE, AND MID CENTURY) FOR 3 DEMAND SCENARIOS ...... 87

v

Executive Summary

vi

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 1

Introduction

Chapter 1 | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 1 - Introduction

The Koshi River Basin profile The Koshi River drains the southern slopes of the Himalayas in Nepal and is formed by three main streams: the Tamur Koshi originating from Mt. Kanchenjunga in the east, Arun Koshi from Mt. Everest in Tibet, and Sun Koshi from Mt. Gosainthan farther west. From their confluence north of the Chatra Gorge onwards, the Koshi River is known as Saptakoshi . The trans-boundary river crosses northern Bihar branching off into several channels and joins the Ganga near Kursela in the Katihar district1.

The eight tributaries of the basin upstream the Chatra Gorge include from east to west2:

• Tamur River with an area of 6,053 km2 (2,337 sq mi) in eastern Nepal; • Arun River with an area of 33,500 km2 (12,900 sq mi), most of which is in Tibet; • Sun Koshi with an area of 4,285 km2 (1,654 sq mi) in Nepal and its northern tributaries Dudh Koshi, Likhu Khola, Tama Koshi, Bhote Koshi and Indravati.

The three major tributaries meet at Triveni, from where they are called Sapta Koshi meaning Seven Rivers. After flowing through the Chatra Gorge the Sapta Koshi is controlled by the Koshi Barrage before it drains into the Gangetic plain.

The total drainage area of the Koshi River is 74,030 km2 out of which 11,410 km2 lies in India and the rest 62,620 km2 lies in Tibet and Nepal3.

Physiography The Koshi River catchment covers six geological and climatic belts varying in altitude from above 8,000 to 95 m comprising the Tibetan plateau, the Himalayas, the Himalayan mid-hill belt, the Mahabharat Range, the Siwalik Hills and the Terai. In the Himalayas, eight peaks over 8,000 m, 36 glaciers and 296 glacier lakes are located in the catchment area4. The Koshi River basin borders the Tsangpo River basin in the north, the Mahananda River basin in the east, the Ganges Basin in the south and the Gandaki River basin in the west5 (Figure 1).

1 http://en.wikipedia.org/wiki/Koshi_River 2 Nepal, S. (2012). Evaluating upstream downstream linkages of Hydrological Dynamics in the Himalayan Region. Dissertation, Faculty of Chemical and Earth Sciences of the Friedrich-Schiller-University Jena 3 http://fmis.bih.nic.in 4 Bajracharya, S. R., Mool, P. K., Shrestha, B. R. (2007). Impact of climate change on Himalayan glaciers and glacial lakes: case studies on GLOF and associated hazards in Nepal and Bhutan. International Centre for Integrated Mountain Development (ICIMOD) 5 Jain, S. K.; Agarwal, P. K.; Singh, V. P. (2007). Hydrology and water resources of India. Dordrecht: Springer. pp. 358–359. ISBN 978-1-4020-5179-1. 8 Chapter 1 - Introduction | INRM Consultants

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 1 : Geographical Context of the Koshi river basin

Of major tributaries, the Arun brings the largest amount of coarse silt in proportion to its total sediment load. The river transports sediment down the steep gradients and narrow gorges in the mountains and foothills where the gradient is at least ten metres per km. On the plains beyond Chatra, the gradient falls below one metre per km to as little as 6 cm per km as the river approaches the Ganga.

Below the Siwaliks, the river has a megafan about 15,000 km2 in areal extent, breaking into more than twelve distinct channels, all with shifting courses due to flooding. Kamla, Baghmati and Budhi Gandak are major tributaries of Koshi in India, besides minor tributaries such as Bhutahi Balan.

Climate The climate in the Koshi River basin ranges from the tropical south to the frigid north. Precipitation is very unevenly distributed throughout the basin. Annual precipitation is about 300–400 mm in the northern Himalayan region, 1000–1500 mm in the subtropical and tropical region and 1500–2500 mm in the temperate region.

The temperature in upper Koshi basin below 00C and maximum is around 100C and in lower Koshi Basin average minimum temperature ranges from 80C to 100C and average maximum maximum temperature ranges from 240C to 250C.

9 Chapter 1 - Introduction | INRM Consultants

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Soils The soils in the basin change from the alluvial soil and upland red loam to the yellow loam in Sun Koshi, then the yellow brown loam in Zhangmu–Nyalam, and the frigid desert soil at the region above 5000 m6. Upper catchment of the Koshi basin consists of Mountain Meadow Soil, Sub- Mountain Meadow Soil, and Brown Hill Soil. The entire lower area of the Koshi Basin in the plains is large inland delta formed by the huge sandy deposit of the Koshi River.

Landuse Upper catchment of the Koshi basin is mountainous and covered with forest and scrubland. Lowe catchment has predominant agricultural landuse. Major crops include paddy, sugarcane, and wheat.

Major Tributaries Important tributaries in Nepal include; Arun Koshi, Tamur Koshi, Sun-Koshi, Indravati, Dudh Koshi, Tama Koshi and Likhukhola (Figure 4). Tributaries in Indian part include Kamla-Balan, Baghmati, Bhuthi Balan and Trijuga on the right side and Fariani dhar, Dhemamadhar on the left side

Irrigation Kosi Barrage, also called Bhimnagar Barrage, was built between 1959 and 1963 and straddles the Indo-Nepal border. It is an irrigation, flood control and hydropower generation project on the Kosi River built under a bilateral agreement between Nepal and India.

6 Chen, N. Sh., Hu, G. Sh., Deng, W., Khanal, N., Zhu, Y. H., and Han, D.: On the water hazards in the trans- boundary Koshi River basin, Nat. Hazards Earth Syst. Sci., 13, 795-808, doi:10.5194/nhess-13-795-2013, 2013 10 Chapter 1 - Introduction | INRM Consultants

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 2

SWAT Model Setup and Current Baseline Simulation

11 Chapter 2 | INRM Consultants.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation

Methodology In the arena of climate change impact assessment deployment of the simulation models is the only option since we are talking of the future which cannot be ascertained through observations. In the present study a hydrological model to simulate the climate change impact on water resources and agriculture has been used. A brief description of the SWAT hydrological model is given in the following paragraphs.

Soil and Water Assessment Tool (SWAT) Model The Soil and Water Assessment Tool (SWAT) model (Arnold et al., 19987, Neitsch et al., 20028) is a distributed parameter and continuous time simulation model. The SWAT model has been developed to predict the hydrological response of un-gauged catchments to natural inputs as well as the manmade interventions. Water and sediment yields can be assessed as well as water quality. The model (a) is physically based; (b) uses readily available inputs; (c) is computationally efficient to operate and (d) is continuous time and capable of simulating long periods for computing the effects of management changes. The major advantage of the SWAT model is that unlike the other conventional conceptual simulation models it does not require much calibration and therefore can be used on un-gauged watersheds (in fact the usual situation).

The SWAT model is a long-term, continuous model for watershed simulation. It operates on a daily time step and is designed to predict the impact of land management practices on water, sediment, and agricultural chemical yields. The model is physically based, computationally efficient, and capable of simulating a high level of spatial details by allowing the watershed to be divided into a large number of sub-watersheds. Major model components include weather, hydrology, soil temperature, plant growth, nutrients, pesticides, and land management. The model has been validated for several watersheds.

In SWAT, a watershed is divided into multiple sub-watersheds, which are then further subdivided into unique soil/land-use characteristics called hydrologic response units (HRUs). The water balance of each HRU in SWAT is represented by four storage volumes: snow, soil profile (0-2m), shallow aquifer (typically 2-20m), and deep aquifer (>20m). Flow generation, sediment yield, and non-point- source loadings from each HRU in a sub-watershed are summed, and the resulting loads are routed

7 Arnold, J. G., R. Srinivasan, R. S. Muttiah, and J. R. Williams. 1998. Large-area hydrologic modeling and assessment: Part I. Model development. J. American Water Res. Assoc. 34(1): 73-89 8 Neitsch, S. L., J. G. Arnold, J. R. Kiniry, J. R. Williams, and K. W. King. 2002a. Soil and Water Assessment Tool - Theoretical Documentation (version 2000). Temple, Texas: Grassland, Soil and Water Research Laboratory, Agricultural Research Service, Blackland Research Center, Texas Agricultural Experiment Station.

12 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

through channels, ponds, and/or reservoirs to the watershed outlet. Hydrologic processes are based on the following water balance equation:

t SWt = SW +∑(Rit − Qi − ETi − Pi − QRi ) i=1 where SW is the soil water content minus the wilting-point water content, and R, Q, ET, P, and QR are the daily amounts (in mm) of precipitation, runoff, evapotranspiration, percolation, and groundwater flow, respectively. The soil profile is subdivided into multiple layers that support soil water processes, including infiltration, evaporation, plant uptake, lateral flow, and percolation to lower layers. The soil percolation component of SWAT uses a storage routing technique to predict flow through each soil layer in the root zone. Downward flow occurs when field capacity of a soil layer is exceeded and the layer below is not saturated. Percolation from the bottom of the soil profile recharges the shallow aquifer. If the temperature in a particular layer is 0ºC or below, no percolation is allowed from that layer. Lateral subsurface flow in the soil profile is calculated simultaneously with percolation. The contribution of groundwater flow to the total stream flow is simulated by routing a shallow aquifer storage component to the stream (Arnold, Allen, and Bernhardt 19939).

SWAT also simulates the nutrient dynamics. Sediment yield is calculated based on the Modified Universal Soil Loss Equation (MUSLE) (Williams, 197510). The movement of nutrients, i.e. nitrogen and phosphorus is based on built in equations for their transformation from one form to the other. The total amounts of nitrates in runoff and subsurface flow is calculated from the volume of water in each pathway with the average concentration. Phosphorus however is assumed to be a relatively less mobile nutrient, with only the top 10 mm of soil considered in estimating the amount of soluble P removed in runoff. A loading function is used to estimate the phosphorus load bound to sediments (McElroy et al, 197611). SWAT calculates the amount of algae, dissolved oxygen and carbonaceous biological oxygen demand (CBOD - the amount of oxygen required to decompose the organic matter transported in surface runoff) entering the main channel with surface runoff. CBOD loading function is based on a relationship given by Thomann and Mueller (1987)12

Advantages of the SWAT model The SWAT model possesses most of the attributes which are identified to be the desirable attributes that a hydrological model should possess.

The SWAT model is a spatially distributed physically based model. It requires site specific information about weather, soil properties, topography, vegetation, and the land management practices being followed in the watershed. The physical processes associated with water movement,

9 Arnold, J.G., Allen, P.M, and Bernhardt, G.T. 1993. A comprehensive surface groundwater flow model. Journal of Hydrology, 142: 47-69 10 Williams, J.R. 1975. Sediment routing for agricultural watersheds. Water Resources Bulletin, 11 (5): 965-974. 11 McElroy, A.D., Chiu, S.Y. and Nebgen, J.W. 1976. Loading functions for assessment of water pollution from nonpoint sources. EPA document 600/2-76-151, USEPA, Athens, GA 12 Thomann, R.V. and J.A. Mueller. 1987. Principles of surface water quality modelling and control. Harper & Row Publishers, New York 13 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

sediment movement, crop growth, nutrient cycling, etc. are directly modelled by SWAT using these input data. This approach results in major advantages, such as:

• Un-gauged watersheds with no monitoring data (e.g. stream gauge data) can be successfully modelled. • The relative impact of alternative input data (e.g. changes in management practices, climate, vegetation, etc.) on water quantity, quality or other variables of interest can be quantified. • The model uses readily available inputs. The minimum data required to make a SWAT run are the commonly available data from local government agencies. • The model is computationally efficient. Simulation of very large basins or a variety of management strategies can be performed without excessive investment of time or money. • The model enables users to study impacts on account of human interventions which makes it very suitable for scenario generation. • The model is also capable of incorporating the climate change conditions to quantify the impacts of change. • The model has gained a wide global acceptability. Currently 720 peer reviewed papers have been published based on the SWAT model (http://swatmodel.tamu.edu ). The current rate of publication is about 120 peer reviewed papers per year. There are more than 90 countries using the model for practical applications and at the least, more than 200 graduate students all over the world are using it as part of their M.S. or Ph.D. research program. In the U.S alone, more than 25 universities have adapted the model in graduate level teaching classes. • SWAT is a public domain model actively supported by the Grassland, Soil and Water Research Laboratory (Temple, TX, USA) of the USDA Agricultural Research Service.

14 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Development of hydrologic model for the Koshi river basin In the present analysis, the part of the Koshi river basin upto Chatra is used. The study area is shown in Figure 2.

Figure 2 : Study area - The Koshi basin

Mapping of a basin on to the SWAT hydrological model involves an elaborate procedure. The following paragraphs briefly describe the methodology used for mapping the Koshi river system contributing to SSP.

Data Used Spatial data and the source of data required for the study area include:

• Digital Elevation Model: SRTM, of 90 m resolution13 • Drainage Network – Hydroshed14 • Soil maps and associated soil characteristics (source: FAO Global soil)15 • Land use: Global Map of Land Use/Land Cover Areas (GMLULCA), IWMI’s Global Map of Irrigated Areas (GMIA) (source: IWMI) 16

13 http://srtm.csi.cgiar.org 14 http://hydrosheds.cr.usgs.gov/ 15 http://www.lib.berkeley.edu/EART/fao.html 15 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

The Hydro-Meteorological data pertaining to the river basin is required for modelling the catchment. These include daily rainfall, maximum and minimum temperature, solar radiation, relative humidity and wind speed. These Weather data available as per following details are used

• Observed rainfall and temperature data provided by ICIMOD and APHRODITE gridded Rainfall data (1971–2006)17. • Princeton University dataset constructed by combining a suite of global observation- based datasets with the NCEP/NCAR reanalysis18 • Climate Change: Global Climate Model outputs, IPCC AR5 BCCR downscaled at a resolution of 0.25° X 0.25° latitude by longitude grid points obtained from ICIMOD for Baseline (1996–2005, BL) , near term (2006-2050, 2041-2050 is considered as MC) for IPCC AR5 RCP 4.5 and RCP 8.5 scenario • Climate Change statistical down scaling outputs from 8 models provided by ICIMOD. Water demand and abstraction data: In the absence of observed data, secondary information shall be used

• Current management/operation practices, existing irrigation as per crop demand taken from the landuse information • Reservoir and diversion locations and characteristics from the published literature.

Model Performance Statistical parameters namely regression coefficients (R2) and Nash Sutcliffe coefficient (NS) are used to assess the model efficiency on monthly SWAT hydrologic stream flow predictions. Stream flow monitoring stations given by ICIMOD are used. Calibration and Validation shall be done only for Koshi and for the period for which data is made available.

Model Evaluation Statistics (Dimensionless)

Nash-Sutcliffe efficiency (NSE): The Nash-Sutcliffe efficiency (NSE) is a normalized statistic that determines the relative magnitude of the residual variance (“noise”) compared to the measured data variance (“information”) (Nash and Sutcliffe, 197019). NSE indicates how well the plot of observed versus simulated data fits the 1:1 line. NSE is computed as

= 푛 표푏푠 푠푖푚 2 ∑푖=1�푌푖 − 푌푖 � 푁푆퐸 � 푛 표푏푠 푚푒푎푛 2� obs th 푖=1 푖 sim th where Yi is the i observation for the constituent∑ �푌 being− 푌evaluated,� Yi is the i simulated value for the constituent being evaluated, Ymean is the mean of observed data for the constituent being

16 http://www.iwmigiam.org/info/main/index.asp 17 http://www.chikyu.ac.jp/precip/products/index.html (IMD gridded data does not exist in the Nepal part, in order to maintain the uniformity of data sources, it is advisable that single source is used wherever possible, and that is the reason Aphrodite is chosen (source of Aphrodite anyway are from individual countrys' Met dept. repository)) 18 http://hydrology.princeton.edu/data/index.html 19 Nash, J. E., and J. V. Sutcliffe. 1970. River flow forecasting through conceptual models: Part 1. A discussion of principles. J. Hydrology 10(3): 282-290 16 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

evaluated, and n is the total number of observations. NSE ranges between−∞ and 1.0 (1 inclusive), with NSE = 1 being the optimal value. Values between 0.0 and 1.0 are generally viewed as acceptable levels of performance, whereas values <0.0 indicates that the mean observed value is a better predictor than the simulated value, which indicates unacceptable performance20

Coefficient of determination (R2): Coefficient of determination (R2) describes the degree of co- linearity between simulated and measured data. R2 describes the proportion of the variance in measured data explained by the model. R2 ranges from 0 to 1, with higher values indicating less error variance, and typically values greater than 0.5 are considered acceptable (Santhi et al., 200121, Van Liew et al., 200322). R2 is oversensitive to high extreme values (outliers) and insensitive to additive and proportional differences between model predictions and measured data (Legates and McCabe, 199923).

Spatial scope/Modelling Scales The analysis is being performed at three spatial scales with the fineness of the input data and analysis of results increases from basin to sub-basin and from sub-basin to catchment levels. The inclusion of infrastructures, their operations and adaptation options shall be worked out in detail for the second level and third level as mentioned below:

a) At first level the whole Ganges Basin is considered b) At the second level Koshi Basin, up to the up to the confluence with the Ganga main trunk. c) At the third level the Tamor, Tama Koshi and Dudh Koshi The following paragraphs describe the hydrological modelling for Koshi, Dudh Koshi, Tama Koshi and Tamor respectively.

20 Moriasi, D. N., J. G. Arnold, M. W. Van Liew, R. L. Bingner, R. D. Harmel, and T. L. Veith, 2007. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations, Transactions of the ASABE, Vol. 50(3): 885−900 2007 21 Santhi, C, J. G. Arnold, J. R. Williams, W. A. Dugas, R. Srinivasan, and L. M. Hauck. 2001. Validation of the SWAT model on a large river basin with point and nonpoint sources. J. American Water Resources Assoc. 37(5): 1169-1188 22 Van Liew, M. W., J. G. Arnold, and J. D. Garbrecht. 2003. Hydrologic simulation on agricultural watersheds: Choosing between two models. Trans. ASAE 46(6): 1539-1551 23 Legates, D. R., and G. J. McCabe. 1999. Evaluating the use of “goodness-of-fit” measures in hydrologic and hydroclimatic model validation. Water Resources Res. 35(1): 233-241 17 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Mapping the Koshi river basin The ArcSWAT interface has been used to pre-process the spatial data for the river system. A digital elevation model (DEM) from SRTM was used for basin delineation and is shown in Figure 3. The DEM has 90 m resolution.

Figure 3 : Digital Elevation Model of the Koshi River basin

The topographic statistics of elevation of the study area is given in Table 1

Table 1 Elevation Summary –Study Area

Parameter Elevation (m Minimum Elevation 8 Maximum Elevation 8672 Mean Elevation 2462

Basin Demarcation Figure 4 shows the automatically delineated catchment with the generated drainage network using the DEM.

18 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 4 : Basin delineation using DEM for the Koshi River basin

Watershed (sub-basin) Delineation Automatic delineation of watersheds was done by using the DEM as input. The target outflow point is interactively selected. The Koshi basin has been delineated using 100,000 hectare as minimum stream threshold and has resulted in 50 sub-basins (Figure 5). Basin area of the Koshi up to the outflow point at Karnataka State boundary is 91,538 sq km.

Figure 5 : Sub basin delineation using DEM for the Koshi River basin

19 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Land Cover/Land Use Layer Land Use/Land Cover is another important segment of data that is required for pre-processing. The landuse , land cover from IWMI, as shown in Figure 6, has been used for the present study. Table 2 gives the land use categories and the area covered under each category of land use for the Koshi basin .

Figure 6 : Landuse map for the Koshi River basin

The major part of the basin is under forest 34%), agriculture land use (30%) with large portion under the irrigated agriculture, and rice, wheat, sugarcane are the predominant crops. Grassland is about 33%.

Table 2 Landuse Categories –of the Koshi river basin

Land Use Area (ha) % of Watershed Area Forest-Deciduous 407699 4 Forest-Evergreen 297901 3 Forest-Mixed 1560744 17 Italian (Annual) Ryegrass 3063791 33 Pine 957708 10 Rice 1861964 20 Corn 122758 1

20 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Land Use Area (ha) % of Watershed Area Sugarcane 700968 8 Tomato 90692 1 Water 78451 1

Soil Layer Information on the soil profile is also required for simulating the hydrological character of the basin. Soil data from Global FAO soil has been used for the modelling. The soil map is shown in Figure 7.

Figure 7 : Soil map for the Koshi River basin

The soil is predominantly clay loam however; sandy clay loam, clay loam are also found. Table 3 shows the area under each of the soil texture class found in the Koshi basin.

Table 3 Soil Texture of the Koshi river basin

Soil Texture - Main Area (ha) % of Watershed Area Loam 8233644 90.0 Glacier 470918 5.2 Clay Loam 411032 4.5 Water 35126 0.4

21 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Hydro-Meteorological and Water resources structures data The daily rainfall data received from ICIMOD for 17 stations (1970-2010) and APHRODITE24 gridded data at a resolution of 0.25° X 0.25° latitude by longitude grid points (1971–2006) has been used. The daily maximum and minimum temperature data received from ICIMOD for 15 stations (1970-2010) and NCEP25 temperature gridded data at a resolution of 0.5° X 0.5° latitude by longitude grid points (1979– 2010) has been used (represented by red dots in Figure 8).

Figure 8 : : Meteorological Locations for the Koshi River basin

Stream Flow Records in Koshi River Basin Flow records of Koshi basin at 7 locations are available. Error! Reference source not found. and Figure 9 give the details of the stations.

Table 4 Stream flow measurement stations –Koshi River Basin

Station No River Name Site Name Latitude Longitude Elevation Period

604.5 Arun River Turkeghat 27.33 87.19 414 1975-2007

610 Bhote Koshi Barhbise 27.79 85.89 840 1965-2007

24 http://www.chikyu.ac.jp/precip/products/index.html 25 The National Centers for Environmental Prediction (NCEP) Climate Forecast System Reanalysis (CFSR), http://globalweather.tamu.edu/ 22 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Station No River Name Site Name Latitude Longitude Elevation Period

630 Sunkoshi River Pachuwarghat 27.56 85.75 589 1964-2010

652 Sunkoshi River Khurkot 27.33 86 455 1968-2009

670 Dudhkoshi River Rabuwa Bazar 27.27 86.66 460 1964-2010

690 Tamor River Mulghat 26.93 87.33 276 1965-1995

695 Saptakoshi River Chatara 26.87 87.16 140 1977-2007

Figure 9 : Stream flow measurement Locations map of Koshi River Basin

Data Limitations and Model Assumptions Assumptions were made due to the following data limitations

• Snow and Glacier: snow and glacier extent and initial glacier depth, depth and spread of glacier, precipitation and temperature lapse rates to represent the orographic effect • Ground Water: initial depth of shallow aquifer • Source and amount of irrigation, cropping pattern taken from secondary information • Soils: In the absence of high resolution soil map and soil profiles, coarse information used

23 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Model limitations • Initialization of various model input parameters such as initial soil moisture, groundwater levels, residue, specific crop growth parameters and related databases has been made. Most of these parameters were initialized by running the model warm-up or setup period for five years to bring the model to an equilibrium stage • Due to the size of the sub-basins the climatic data were interpolated to assign one weather station per sub-basin, where the model assumes uniform input for the entire sub-basin. However the weather is very dynamic in spatial and temporal scales. This could add to uncertainties to model inputs and thus model predictions and outputs • The snow process simulations are very complex and highly spatially variable due to slope, depth of snow, aspect, location of snow line and initialization of glacier depth • The specific sources, amounts and frequencies of irrigation for individual crops are not known throughout the basin, hence data available from various sources were compiled and used as average irrigation schedules across the basin.

SWAT Model Performance for the Koshi Basin Although the SWAT model does not require elaborate calibration (Gosain et al., 200526 ), model validation has been made using the observed data for the period 1964-2006 at monthly scale. Data for 7 Stream flow monitoring stations were made available with daily interval for various periods ranging from 1964 to 2010 (Figure 9). However, model comparisons were performed at monthly scale due to uncertaintes in other data as described in the previous paragraphs. Before performing any of the statistical comparison, the model was setup with best of the inputs available and the results from the model for some of the water balance components such as general evapotranspiration, runoff and base flow/return flow were analyzed. Since most of the stream flow data were available in the Nepal-region, it captures most of the snow hydrology components with limited or no abstraction or structures in this area. There is very limited agriculture or any significant landuse with consumptive water use above these gauge locations. SWAT model was setup with elevation bands, temperature and precipitation lapse rates along with assumed glacier depth. Statistical parameters namely regression coefficients (R2) and Nash Sutcliffe coefficient (NS) were used to assess the model efficiency of simulated monthly stream flows.

The long-term simulated monthly mean at all drainage areas outlets are with observed means, the R2 and Coefficient of efficiency are above literature acceptable ranges27. Table 5 shows the model performance parameters. As it can be seen that the SWAT model as setup in this study with elevation bands, temperature and precipitation lapse rates along with initial glacier depth produced statistically satisfactory results in comparison with observed data at various locations.

26 Gosain, A.K., Sandhya Rao, Srinivasan, R. and Gopal Reddy, N., 2005. “Return-Flow Assessment for Irrigation Command in the Palleru River Basin Using SWAT Model”. Hydrological Processess 19, 673-682. 27 Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Bingner, R.L., Harmel, R.D., Veith, T.L. 2007. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Transactions of the ASABE. 50(3):885-900 24 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Table 5 SWAT output comparison Locations and model efficiency parameters

Gauge Site River Catchment Mean Start End COE** Correlation RSR+ PBIAS++ Area*(Sq km) Flow* Year Year coefficient (%) Barhbise Bhote NA (2527) 80.6 1965 1995 0.66 0.17 0.34 9.9 (175) Pachuwarghat Sunkoshi NA (4978) 187.6 2001 2006 0.85 0.96 0.15 -20.4 (225.9) Khurkot Sunkoshi NA (10484) 490 1968 1980 0.63 0.88 0.37 -6.7 (477.3) Rabua Bazar Dudkoshi NA (3842) 201.1 1964 1998 0.75 0.89 0.25 -16.0 (233.4) Turkeghat Arun NA (29883) 459.9 1975 2006 0.72 0.91 0.29 -23.1 (566.3) Mulghat Tamor NA (6128) 311.4 1965 1995 0.67 0.83 0.33 -10.9 (345.3) Chatara Saptakoshi NA (57326) 1530.4 1977 2006 0.71 0.88 0.30 -18.3 (1809.8) * Model parameter is shown in bracket, ** Nash-Sutcliffe coefficient, + Ratio of the root mean square error to the standard deviation of measured data (RSR), ++ Volume Bias, NA: Not available

The simulated mean on long-term basis at all drainage area levels were on par with observed means, wherein the R2 and Coefficient of efficiency were above literature acceptable ranges of 0.79 to 0.93 and 0.56 to 0.91 respectively.

The elevation range in Koshi basin is from 8000m to 2000 m at the foot hills of Himalayas. Hence, all the sub-basins are above 2000 m elevation, and corresponding area that fall within each of the elevation bands were incorporated in the model sub-basin input files. A literature based value of -2 to -4oC/km change28 was used as temperature lapse rate and 100 mm/km was used as precipitation lapse rate for those sub-basins where the elevation bands were incorporated. These correction factors are necessary to account for change in precipitation and temperature at higher altitudes since the observation of rainfall and temperatures were very limited and that too confined to lower elevations. In addition, there were no data that depicts the spatial pattern of glacier depth and snow pack information. In this modelling setup it is assumed that there is glacier depth of 35 m to 50 m for altitudes about 4500 m elevation29. The model provides the variation in glacier depth over time to account for the loss of glacier due to climatic factors including climate change. The method deployed here is also another possible way to estimate the volume of glacier and snow (in the absence of actual observations) to generate the simulated stream flow at various locations. Figure 10 provides the results of calibration for various locations in the Koshi basin in the form of time series plots as well as scatter plots.

28 snow and glacier hydrology,. Pratap Singh and Vijay P. Singh. Kluwer Academic Publishers, 2001 29 Ganga Glacier data by ICIMOD: Ganga_Gr_Final.shp 25 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 10 : SWAT output comparison Locations and model efficiency parameters for the Koshi basin

Observed Vs. Simulated Time Series Plots Observed Vs. Simulated Scatter Plots Barhbise on Bhote river

Pachuwarghat on Sunkoshi river

Khurkot on Sunkoshi river

Rabua Bazar on Dudkoshi river

26 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 10 : SWAT output comparison Locations and model efficiency parameters for the Koshi basin

Observed Vs. Simulated Time Series Plots Observed Vs. Simulated Scatter Plots Turkeghat on Arun river

Mulghat on Tamor river

Chatara on Saptakoshi river

The Hydrologic Simulation with Observed Weather Following model setup, it has been validated using the available stream flow observation stations. After setting up the model for the baseline observed weather, simulation runs were made for a period of 29 years (1970 to 2010) with 4 years model warm-up period. The model has been run on continuous daily basis for all the sub-basins of the Koshi basin .

The outputs provided by the model are very exhaustive covering all the components of water balance spatially and temporally. The sub components of the water balance that are more significant and were used for analyses include: 27 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

• Precipitation • Total flow (Water yield) consisting of surface runoff, lateral flow and base flow • Actual evapotranspiration (Actual ET) • Base flow • Ground water recharge.

The outputs can be depicted in many ways depending on the focus and requirement of the user. These components are expressed in terms of total annual depth of water in mm over the total watershed area. In other words, the total water yield is the equivalent depth in mm, of flow past the outlet of the watershed on average annual and monthly basis.

Temporal distribution of major water balance components Figure 11 depicts the average annual precipitation, water yield, evapotranspiration and ground water recharge for the entire Koshi River Basin.

Figure 11 : Simulated Average water balance components for Koshi River Basin

2500 Koshi Basin - Yearly Water Balance

2000

1500

Value (mm) Value 1000

500

0

1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Rainfall Water Yield Evapotranspiration Groundwater Recharge

As shown in Figure 11, inter annual variation of rainfall for the period of analysis from 1974-2010 was 1860 to 2486 mm, while inter annual variation of water yield was 928 to 1493 mm for the same period. This shows that there is some degree of fluctuations in the rainfall and runoff pattern even in Koshi basin.

28 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

The Figure 12 depicts the long term monthly (1974-2010) average water balance components of rainfall, water yield and evapotranspiration for Koshi basin. It can be seen that the rainfall contribution to water yield and evapotranspiration is mainly in the months of July, August and September with a maximum in August while for other months the contribution is from snow melt. In August about 69% of the rainfall is converted into water yield and 14% of the rainfall contributes to evapotranspiration. The remaining has been converted to ground water storage, soil moisture build up, etc.

Figure 12 : Simulated Average monthly water balance components for Koshi River Basin

475 450 425 400 375 350 325 300

275 250 225

Value (mm) Value 200 175 150 125 100 75 50 25 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months Rainfall Snowfall Water Yield Evapotranspiration Koshi Basin - Monthly Water Balance

The Figure 13 shows long-term annual water balance components and their contribution in rainfall. It can be seen from the figure that between ET and ground water contribution, almost 34% of the available water is utilized leaving about 69% as streamflow in the river resulted due to rainfall in the Koshi basin.

29 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 13 : Simulated Average annual water balance components for Koshi River Basin

Value (mm)

0 200 400 600 800 1000 1200 1400 1600

Precipitation 1409.0

Snowfall 70.0

Snowmelt 106.0

WaterYield 971.0 (69%)

Baseflow 506 (36%)

Evapo-transpiration 347 (25%)

Groundwater Recharge 131.0 (9%) Koshi Basin - Annual Water Balance

Precipitation Snowfall Snowmelt WaterYield

% contribution in rainfall is shown in brackets

Flow Duration Curve Assessment of dependable lean season flows along with their distribution in time is essential for planning and development of water supply schemes. The dependability of the water yield of the Koshi basin has been analyzed and shown in Figure 14. Flow dependability for various sub-basins of Koshi with respect to four levels (25, 50, 75 and 90%) of dependability is shown in Table 6. The 90% dependability level is considered safe for determining assured water supply.

Table 6 Flow Dependability – Sub-basins of Koshi River Basin

Dependability Barhbise Pachuwar Turkeghat Khurkot Rabuwa Mulghat Chatara (%) Ghat Bazar 25 145 526 1038 917 358 577 3473 50 70 180 419 365 128 276 1315 75 32 66 127 123 46 113 530 90 21 40 60 69 22 68 276 Flow units: cumecs

30 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 14 : Flow Duration Curves at various locations on Koshi River Basin

800

700

600

500

400

Discharge (cms) Discharge 300

200

100

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Barhbise

1400

1200

1000

800

600 Discharge (cms) Discharge 400

200

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Pachuwar Ghat

31 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 14 : Flow Duration Curves at various locations on Koshi River Basin

2500

2000

1500

1000 Discharge (cms) Discharge

500

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Khurkot

1000

900

800

700

600

500

400 Discharge (cms) Discharge 300

200

100

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Rabuwa Bazar

32 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 14 : Flow Duration Curves at various locations on Koshi River Basin

7000

6000

5000

4000

3000 Discharge (cms) Discharge 2000

1000

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Turkeghat

1200

1000

800

600

Discharge (cms) Discharge 400

200

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Mulghat

33 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 14 : Flow Duration Curves at various locations on Koshi River Basin

10000

9000

8000

7000

6000

5000

4000 Discharge (cms) Discharge 3000

2000

1000

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 % of times flow equalus or exceeds

Chatara

34 Chapter 2 - SWAT Model setup for the Koshi basin and Current Baseline Simulation | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 3

Climate Change Scenario Simulation (BCCR Dynamic Downscale)

35 Chapter 3 | IIT Delhi

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data

The IPCC scenarios provide a mechanism to assess the potential impacts on climate change. Global emission scenarios were first developed by the IPCC in 1992 and were used in global general circulation models (GCMs) to provide estimates for the full suite of greenhouse gases and their potential impacts on climate change. Since then, there has been greater understanding of possible future greenhouse gas emissions and climate change as well as considerable improvements in the general circulation models. The IPCC, therefore, developed a new set of emissions scenarios. The process by which these new scenarios are being produced differs from earlier scenario development.

The new process aims to both shorten the time required to develop and apply new scenarios, and to ensure better integration between socio-economic driving forces, changes in the climate system, and the vulnerability of natural and human systems. Rather than starting with socio-economic scenarios that give rise to alternative greenhouse gas emissions, the new scenarios take alternative futures in global greenhouse gas and aerosol concentrations as their starting point. These are called Representative Concentration Pathways (RCPs)30. The Representative Concentration Pathways (RCP) are based on selected scenarios from four modelling teams/models working on integrated assessment modelling, climate modelling, and modelling and analysis of impacts.

RCPs are four greenhouse gas trajectories adopted by the IPCC for its fifth Assessment Report (AR5). The four RCPs; RCP2.6, RCP4.5, RCP6, and RCP8.5, are named after a possible range of radiative forcing values in the year 2100. Table 7 gives the overwiew of four RCPs.

Table 7: Overview of Representative Concentration Pathways (RCPs) adopted by IPCC AR5

RCP Description IA Model Publication – IA Model RCP8.5 Rising radiative forcing pathway leading to MESSAGE Riahi et al. (2007), Rao 8.5 W/m2 in 2100. & Riahi (2006) RCP6 Stabilization without overshoot pathway AIM Fujino et al. (2006), to 6 W/m2 at stabilization after 2100 Hijioka et al. (2008) RCP4.5 Stabilization without overshoot pathway GCAM (MiniCAM) Smith and Wigley to 4.5 W/m2 at stabilization after 2100 (2006), Clarke et al. (2007), Wise et al. (2009) RCP2.6 Peak in radiative forcing at ~ 3 W/m2 IMAGE van Vuuren et al. before 2100 and decline (2006; 2007) Source: http://sedac.ipcc-data.org/ddc/ar5_scenario_process/RCPs.html

Climate change Data Used for Hydrological Modelling Data for many different variables (physical quantities, Rainfall, Temperature, Solar Radiation, Relative humidity, Wind speed) at a variety of different timescales; Daily, IPCC AR5 BCCR downscaled at a resolution of 0.25° X 0.25° latitude by longitude grid points obtained from ICIMOD for Baseline

30 http://sedac.ipcc-data.org/ddc/ar5_scenario_process/index.html 36 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

(1996–2005, BL) , near term (2006-2050, 2041-2050 is considered as MC) for IPCC AR5 RCP 4.5 and RCP 8.5 scenario, extracted for the relevant grids falling in the Koshi river basin (Figure 15) for the time periods Baseline (1996 - 2005), Mid Century (2041 – 2050) has been used.

Figure 15 :Climate Change Grid Locations in the Koshi River basin

The Hydrologic Simulation with Climate Change Scenarios The model has been run using climate scenarios for near term (MC, 2006 – 2051) without changing the land use. The climate change simulations are run for 2 RCP scenarios (RCP 4.5 and RCP 8.5), for 2 time periods; Baseline and Mid Century.

The outputs of these scenarios have been analyzed to evaluate the possible impacts on the runoff, snowmelt, baseflow, soil moisture, ground water recharge and actual evapotranspiration (expressed as change between the baseline and future periods).

Figure 16 and Figure 17 present the snapshot long-term variability of the key water balance elements for the Koshi basin as a single unit. These components are expressed in terms of total annual depth of water in mm over the total watershed area. In other words, the total water yield is the equivalent depth in mm, of flow past the outlet of the watershed on average annual basis.

37 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 16 Annual water balance components for BL and MC climate scenarios (IPCC AR5 RCP 4.5) for the Koshi River basin

Annual Value (mm) -500 0 500 1000 1500 2000 2500

Baseline (1998-2005) 2107.8

Rainfall Mid Century (2041-2050) 2258.8

Baseline (1998-2005) 1591.9

Mid Century (2041-2050) 1586.3 Water Yield

Baseline (1998-2005) 326.3

Snowmelt Mid Century (2041-2050) 154.6

Baseline (1998-2005) 343.5

on Mid Century (2041-2050) 377.3 Evapotranspirati

Baseline (1998-2005) 863.9

Baseflow Mid Century (2041-2050) 784.8

Baseline (1998-2005) -338.3

Recharge Mid Century (2041-2050) 208.3 Groundwater

Precipitation Water Yield Snowmelt Evapotranspiration Baseflow Recharge

Koshi Basin : RCP 4.5

38 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 17 Annual water balance components for BL and MC climate scenarios (IPCC AR5 RCP 8.5) for the Koshi River basin

Annual Value (mm) 0 500 1000 1500 2000 2500

Baseline (1998-2005) 2107.8 Rainfall Mid Century (2041-2050) 2233.7

Baseline (1998-2005) 1591.9

Water Yield Mid Century (2041-2050) 1581.6

Baseline (1998-2005) 326.3

Snowmelt Mid Century (2041-2050) 134.5

Baseline (1998-2005) 343.5

n

Mid Century (2041-2050) 364.2 Evapotranspiratio

Baseline (1998-2005) 863.9

Baseflow Mid Century (2041-2050) 790.1

Baseline (1998-2005) 214.8

Recharge Mid Century (2041-2050) 208.4 Groundwater

Precipitation Water Yield Snowmelt Evapotranspiration Baseflow Recharge Koshi Basin : RCP 8.5

The results of hydrological simulation have been depicted in Figure 18 in terms of long term monthly average by picking up five major elements of water balance i.e., precipitation, water yield, snowmelt, evapotranspiration and ground water recharge.

39 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 18 Monthly water balance components for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for the Koshi River basin

600 Koshi Basin 550 500 450 400

350 300 250 200 Value (mm) 150 100 50

0 Jul Jul Jul Jan Jan Jan Jun Jun Jun Oct Oct Oct Apr Apr Apr Feb Feb Feb Sep Sep Sep Dec Dec Dec Aug Aug Aug Nov Nov Nov Mar Mar Mar May May May Baseline (1996-2005) Mid Century (2041-2050) Mid Century (2041-2050) RCP 4.5 RCP 8.5 Rainfall Water Yield Snowmelt Evapotranspiration Groundwater Recharge

The results of hydrological simulation have also been depicted spatially in Figure 16 in terms of long term annual average by picking up three major elements of water balance i.e., precipitation, evapotranspiration and water yield.

40 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 19 Spatial distribution of change in water balance components towards mid century (IPCC AR5 RCP 4.5 and 8.5) for the Koshi River basin

RCP 4.5

RCP 8.5

41 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Seasonal Change in water balance components for RCP 4.5 and RCP 8.5 Scenarios Longterm average of water balance components over baseline (1996-2005) and mid century (2041- 2050) scenarios have been used for assessing change from baseline to mid century. Table 8 provides the summary of change in water balance components as percentage of the change in precipitation from baseline to mid century respectively for RCP 4.5 and RCP 8.5 scenarios.

Table 8 Summary of Seasonal Change in water balance (mm) expressed as percent change in precipitation for the study area (baseline to mid century) – RCP 4.5 and RCP 8.5

Scenario/Season Precipitation Evapo- Surface Baseflow+ Total Ground transpiration Runoff Water water Yield* Recharge** RCP 4.5 - MC Southwest Monsoon (June to September - JJAS) Avg JJAS (Baseline) mm 1523.6 156.6 373.9 499.0 1122.0 398.4 Avg JJAS (Mid Century) mm 1602.4 174.6 410.1 461.9 1105.1 343.9 Percent Change in 5.2 precipitation Net change (mm) 78.8 18.0 36.2 -37.1 -16.8 -54.4 Change (%)*** 22.8 46.0 -47.2 -21.4 -69.1 Northeast Monsoon (October to December - OND) Avg OND (Baseline) mm 192.8 65.8 26.4 269.1 314.7 -211.7 Avg OND (Mid Century) mm 304.0 70.7 75.4 245.2 344.9 -167.8 Percent Change in 57.7 precipitation Net change (mm) 111.2 4.9 49.0 -23.9 30.2 43.9 Change (%)*** 4.4 44.1 -21.5 27.2 39.5 RCP 8.5 - MC Southwest Monsoon (June to September - JJAS) Avg JJAS (Baseline) mm 1523.6 156.6 373.9 499.0 1122.0 398.4 Avg JJAS (Mid Century) mm 1629.1 166.2 428.1 464.3 1123.4 345.4 Percent Change in 6.9 precipitation Net change (mm) 105.5 9.6 54.2 -34.7 1.4 -53.0 Change (%)*** 9.1 51.3 -32.9 1.3 -50.2 Northeast Monsoon (October to December - OND) Avg OND (Baseline) mm 192.8 65.8 26.4 269.1 314.7 -211.7 Avg OND (Mid Century) mm 257.6 68.9 54.1 246.6 325.3 -167.0 Percent Change in 33.6 precipitation Net change (mm) 64.8 3.1 27.7 -22.5 10.5 44.7 Change (%)*** 4.8 42.8 -34.8 16.2 69.0 +: Baseflow: contributes to stream flow during non rainy period * Water Yield (streamflow): surface runoff+baseflow+lateral flow ** Groundwater Recharge: shallow and deep aquifer recharge ***Distribution of water balance components as percentage of change in precipitation

42 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

It can be seen from Table 8 that projected change in precipitation is higher during north east monsoon (RCP 4.5 - 58%, RCP 8.5 - 34%) as compared to the south west monsoon period (RCP 4.5 - 5%, RCP 8.5 - 7%). Most of the increase in rainfall get reflected in increase in evapotranspiration and direct surface runoff. RCP 4.5 shows decrease in water yield with higher evapotranspiration as compared to RCP 8.5 (marginal increase in water yield).

The change in simulated surface water runoff, groundwater recharge and evapotranspiration from the baseline to mid century are shown in Figure 20. The RCP 4.5 scenario indicates an increase in south west monsoon (JJAS) precipitation in the Koshi basin by about 5% (79 mm) by mid century. The model results indicate that around 46 % of this increase in precipitation will contribute to increase in runoff. Increase in precipitation in north east monsoon (OND) is higher at 58% (111 mm) towards mid century. Around 44% of this will go as direct surface runoff which contributes to stream flow. There is a marginal increase in evapotranspiration.

The RCP 8.5 scenario indicates an increase in south west monsoon (JJAS) precipitation in the Koshi basin of about 7% (105 mm) by mid century. The model results indicate that around 51 % of this increase in precipitation will get converted to runoff. Increase in precipitation in north east monsoon (OND) is 34 % (65 mm) towards mid century. Around 43% of this will go as direct surface runoff.

Figure 20 Seasonal Change in water balance (mm) expressed as percent change in precipitation (IPCC AR5 RCP 4.5 and RCP 8.5) for the Koshi River basin

43 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Impact Assessment The outputs from the hydrological model have been used to assess the impact of the climate change on the river basins in terms of occurrence of droughts and floods. The rainfall, runoff and actual evapotranspiration have been selected from the available model output since they mainly govern the two extreme impacts due to climate change, namely droughts and floods.

Drought Analysis Drought indices are widely used for the assessment of drought severity by indicating relative dryness or wetness affecting water sensitive economies.

The Palmer Drought Severity Index (PDSI) is one such widely used index that incorporates information on rainfall, land-use, and soil properties in a lumped manner (Palmer 196531). The Palmer index categorize drought into different classes. PDSI value below 0.0 indicates the beginning of drought situation and with a value below -3.0 as sever drought condition.

Soil moisture index developed (Narasimhan and Srinivasan, 200532) to monitor drought severity using SWAT output to incorporate the spatial variability has been used in the present study to focus on the agricultural drought where severity implies cumulative water deficiency. Weekly information has been derived using daily SWAT outputs which in turn have been used for subsequent analysis of drought severity.

The severity of drought effect is proportional to the relative change in climate. For example, if a climate that usually has very slight deviation from the normal experiences a moderate dry period, the effect would be quite dramatic. On the other hand, a very dry period would be needed in a climate that is used to large variations to produce equally dramatic effects. In the current context scale 1 (Index between 0 to -1) represent the drought developing stage and scales 2 (Index between -1 to -4) represent mild to moderate and extreme drought condition.

Soil Moisture Deficit Index (SMDI) was calculated from simulated soil moisture data for baseline (1996-2005) and MC (2041-2050) climate change scenarios. The change from current condition to mid century shows improvement in the drought onset conditions (onset of drought) for Mid Century scenario under IPCC AR5 RCP 4.5. The areas which may fall under moderate to extreme drought conditions (drought index value between -1 to -4) during south west monsoon period show the increase in severity of drought from baseline to Mid Century scenario in RCP 4.5 scenario. The trend is similar in IPCC AR5 RCP 8.5 scenario.

The spatial distribution of change in number of weeks (baseline to mid century) in drought weeks during south west monsoon period are also shown using the SWAT output for smaller drainage basins in the GIS format in Figure 21. The hotspot areas can be recognised from the figure.

31 Palmer, W.C., 1965. Meteorological drought. Research Paper 45. U.S. Department of Commerce, Weather Bureau, Washington, D.C. 58pp. 32 Narasimhan, B. and Srinivasan, R., 2005. Development and evaluation of Soil Moisture Deficit Index (SMDI) and Evapotranspiration Deficit Index (ETDI) for agricultural drought monitoring, Agricultural and Forest Meteorology 133 (2005) 69–88 44 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 21 Change in Spatial distribution of Soil moisture deficit weeks for BL and MC scenarios (IPCC AR5 RCP 4.5) for the Koshi River basin – South West Monsoon (JJAS)

RCP 4.5

45 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 22 shows distribution of drought weeks for IPCC AR5 RCP 8.5 scenario. It can be seen that the drought situation during south west monsoon period improves towards Mid Century when compared to Baseline for drought onset condition. However, the conditions are expected to deteriorate especially in the northern part of the Koshi basin as far as moderate to extreme drought conditions are concerned.

Figure 22 Change in Spatial distribution of Soil moisture deficit weeks for BL and MC scenarios (IPCC AR5 RCP 8.5) for the Koshi River basin – South West Monsoon (JJAS)

RCP 8.5

IPCC AR5 RCP 4.5 shows high water stress in many parts of the basin than RCP 8.5 scenario.

The concept of drought week is reflective of the change in the normal moisture condition of a location. In this sense, if a dry or desert area is analysed which has uniform conditions over a long period, it shall declare all the weeks to be normal weeks since there is no change in the character of the area on the basis of the conditions prevalent in the specific week of the year over a long period e.g., 30 years. Variability in moisture condition is expected to increase for MC and stabilize in EC scenario. An annual increase of about 2 to 3 weeks under drought is expected in mid century.

Flood Analysis The vulnerability assessment with respect to the possible future floods has been carried out using the daily outflow discharge taken for each sub-basin from the SWAT output. These discharges have been analysed with respect to the maximum annual peaks. Maximum daily peak discharge has been

46 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

identified for each year and for each sub-basin. Analysis has been performed to identify those basins where flooding conditions may deteriorate under the climate change scenario. Two kinds of analysis has been performed, (i) change in the magnitude of flood peaks above 99th percentile flow has been evaluated and (ii) flow duration curves have been plotted for the Koshi river basin for baseline (1961-1990) and MC (2021-2050) based on the simulated annual flows at various locations.

Average peak discharge at 99th percentile Figure 23 shows peak discharge equal to or exceeding at 1 % frequency from baseline to MC scenarios for Koshi river basin .

Figure 23 Spatial variation in Change in stream discharge at 99th percentile for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for the Koshi River basin

It can be seen from the figure that the magnitude of peak discharge increases in RCP 4.5 mid century scenario as compared to the RCP 8.5 mid century scenario. Projected peak discharge is higher in RCP 4.5 scenario.

Flow Duration Curve and Impact on Reservoirs Assessment of dependable lean season flows along with their distribution in time is essential for planning and development of water supply schemes. The impact of the climate change on the dependability of the water yield of the Koshi river system has been analyzed with respect to three levels of 50, 75 and 90% dependability. The 90% exceedance probability level is considered safe for

47 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

determining assured water supply. Figure 24 shows the locations where flow dependability is assessed for both RCP 4.5 and RCP 8.5 scenario towards mid century.

Figure 24 Locations where Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for the Koshi River basin

Impact of climate change on the dependable flow at various locations shown in Figure 24 is given in Table 9 in million cubic meters. With the available climate change projections it can be inferred that the supply at these locations may not have much reduction at the 75% dependable flow level for IPCC RCP 4.5. However for IPCC AR5 RCP 8.5 scenario marginal reduction in flow dependability at 75% and 90% is projected towards mid century, at many locations.

48 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 25 Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for seven locations in Koshi River basin

500 450 Barhbise on Bhote Koshi 400

350 300 250 200 150

Stream Flow (cumecs) 100 50 0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow 7,000 Pachuwar Ghat on Sun Koshi 6,000

5,000

4,000

3,000

2,000 Stream Flow (cumecs) 1,000

0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow

49 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 25 Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for seven locations in Koshi River basin

14,000 Turkeghat on Arun 12,000

10,000

8,000

6,000

4,000 Stream Flow (cumecs) 2,000

0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow

14,000 Khurkot on Sun Koshi 12,000

10,000

8,000

6,000

4,000 Stream Flow (cumecs) 2,000

0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow

50 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 25 Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for seven locations in Koshi River basin

45,000

40,000 Rabuwa Bazar on Dud Koshi

35,000

30,000

25,000

20,000

15,000

10,000 Stream Flow (cumecs)

5,000

0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow

18,000

16,000 Mulghat on Tamor

14,000

12,000

10,000

8,000

6,000

4,000 Stream Flow (cumecs)

2,000

0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow

51 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 25 Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for seven locations in Koshi River basin

140,000 Chatara on Sun Koshi 120,000

100,000

80,000

60,000

40,000 Stream Flow (cumecs) 20,000

0 0 10 20 30 40 50 60 70 80 90 100 % of times flow equals or exceeds Baseline RCP 4.5 RCP 8.5

Dependability using Daily Stream Flow

Table 9 Dependable flow at various locations (MCM) for IPCC AR5 RCP 4,5 and RCP 8.5 scenarios (Baseline, and mid century)

RCP 4.5 RCP 8.5 Location River Baseline Mid Century Mid Century 25 % Dependable Flow (MCM) Barhbise Bhote 1895 2762 2739 Pachuwar Ghat Sunkoshi 8525 9869 9852 Khurkot Sunkoshi 18331 19264 19287 Rabuwa Bazar Dudkoshi 3931 4943 4587 Turkeghat Arun 47084 39219 38235 Mulghat Tamor 15527 20510 18964 Chatara Saptakoshi 97676 97364 95304 50 % Dependable Flow (MCM) Barhbise Bhote 1734 2657 2621 Pachuwar Ghat Sunkoshi 7839 9389 9265 Khurkot Sunkoshi 17097 18114 18671 Rabuwa Bazar Dudkoshi 3125 4301 3982 Turkeghat Arun 42004 36046 36477 Mulghat Tamor 14686 18383 17641 Chatara Saptakoshi 91462 94530 93735 75 % Dependable Flow (MCM) Barhbise Bhote 1662 2438 2424

52 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

RCP 4.5 RCP 8.5 Location River Baseline Mid Century Mid Century Pachuwar Ghat Sunkoshi 7591 7901 8253 Khurkot Sunkoshi 15841 14897 15286 Rabuwa Bazar Dudkoshi 2825 3380 2926 Turkeghat Arun 35695 34788 34813 Mulghat Tamor 13719 14614 14021 Chatara Saptakoshi 80270 82088 77590 90 % Dependable Flow (MCM) Barhbise Bhote 1606 2278 2371 Pachuwar Ghat Sunkoshi 6938 6684 6601 Khurkot Sunkoshi 13927 12887 12863 Rabuwa Bazar Dudkoshi 2533 2425 2350 Turkeghat Arun 34305 34199 34240 Mulghat Tamor 12559 11444 11228 Chatara Saptakoshi 76048 73348 72926

Limitations of the Study

Assumptions A few assumptions were made for running the scenarios using the hydrological model for the study area. The following major decisions/assumptions have been made while formulating these scenarios:

• Current crop management practices (irrigation from Surface and Ground water) based on landuse map, irrigation source map, command area map and district-wise average irrigation (by source) information was used • No change in future landuse.

Data Limitations Lack of detailed data on:

• Soils: Absence of high resolution Indian soil map and soil profiles • Landuse: Absence of high resolution Indian landuse specially the cropping pattern map

Weather data:

• actual observed station data on rainfall, temperature at daily scale, absence of high altitude weather station locations, spatial distribution issues to address the spatial variability of weather parameters • use of weather generator for generating weather parameters of wind speed, relative humidity etc. • Ground Water: initial depth of shallow and deep aquifer • Irrigation assumptions on the source and amount of irrigation, cropping pattern • Uncertainties in the climate projection

53 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

• Single model climate change simulation at daily scale.

Model limitations • Initialization of various model input parameters such as initial soil nutrient levels, soil moisture, groundwater levels, residue, specific crop growth parameters and related databases have uncertainties. Most of these parameters were initialized by running the model warm-up or setup period for four years to bring the model to an equilibrium stage. • The exact specific irrigation sources, amounts and frequencies for specific crop is not known throughout the basin, hence data available from various sources were compiled and used as average irrigation schedules across the basin. • Fertilizer and manure applications were also quite different over the space, due to lack of information SWAT model automatic fertilizer option was used in some of the subbasins and crops, then in the other areas where data were available the typical application date and amount were used.

54 Chapter 3 - Climate Change Scenarios using BCCR IPCC AR5 RCP weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 4

Climate Change Scenario Simulation - Delta Change Method

55 Chapter 4 | IIT Delhi

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data

Climate change Data Used for Hydrological Modelling Data provided by ICIMOD (Future Water) has been used for the study. GCM Models used for downscaling using delta method are given in Table 10.

Table 10 Model selection based on 10th and 90th percentile values of projected changes in P and T from 1961- 1990 to 2021-2050

Description RCP dP (%) dT (K) Selected Model DRY, COLD RCP45 -1.8 1.4 GISS-E2-R-r4i1p1_rcp45 DRY, WARM RCP45 -1.8 2.3 IPSL-CM5A-LR-r4i1p1_rcp45 WET, COLD RCP45 8.9 1.4 CCSM4-r5i1p1_rcp45 WET, WARM RCP45 8.9 2.3 CanESM2-r4i1p1_rcp45 DRY, COLD RCP85 -1.1 1.7 GFDL-ESM2G-r1i1p1_rcp85 DRY, WARM RCP85 -1.1 2.7 IPSL-CM5A-LR-r4i1p1_rcp85 WET, COLD RCP85 12.1 1.7 CSIRO-Mk3-6-0-r3i1p1_rcp85 WET, WARM RCP85 12.1 2.7 CanESM2-r4i1p1_rcp85

12 grids with monthly delta change data for the future (2021-2050) relative to a reference period (1961-1990) to reflect the change in temperature (K) and precipitation (%) over 60 years was provided. Transient time series was generated using the monthly delta change grids using the observed rainfall (station as well as gridded Aphrodite) and temperature33.

The Hydrologic Simulation with Climate Change Scenarios The SWAT hydrological model has been run using climate scenarios for near term (MC, 2006 – 2051) without changing the land use. The simulations under climate change are run for 2 RCP scenarios (RCP 4.5 and RCP 8.5).

The outputs of these scenarios have been analyzed to evaluate the possible impacts on the runoff, snowmelt, baseflow, soil moisture, ground water recharge and actual evapotranspiration (expressed as change between the baseline and future periods).

Flow Duration Curve and Impact on dependable flow for Koshi Basin Assessment of dependable lean season flow along with their distribution in time is essential for planning and development of water supply schemes. The impact of the climate change on the dependability of the water yield of the Koshi river system has been analyzed with respect to three

33 Manual to generate downscaled climate change scenarios for the downstream basins of the HICAP project, 2012, W.W. Immerzeel 56 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

levels of 50, 75 and 90% dependability. The 90% exceedance probability level is considered safe for determining assured water supply. Figure 24 shows the locations where flow dependability is assessed for both RCP 4.5 and RCP 8.5 scenario towards mid century.

Figure 26 Locations where Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for the Koshi River basin

Impact of climate change on flow at various locations shown in Figure 24 is given in Table 9 for various dependable levels. With the available climate change projections it can be inferred that the flow availability at these locations may not have much reduction at the 75% dependable flow level for IPCC RCP 4.5 and RCP 8.5 scenarios. However some locations, like Rabua Bazar, Mulghat and Turkeghat, show marginal decrease at various dependability levels of 50%, 75% and 90%.

57 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 27 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

700 Barhbise 600

500

400

300 Stream Flow CMS 200

100

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5 1000

900 Pachuwar Ghat 800

700

600

500

400

Stream Flow CMS 300

200

100

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

58 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 27 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

1800

1600 Khurkot

1400

1200

1000

800

Stream Flow CMS 600

400

200

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5 700 Rabuwa Bazar 600

500

400

300 Stream Flow CMS 200

100

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

59 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 27 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

1800 Turkeghat 1600

1400

1200

1000

800

Stream Flow CMS 600

400

200

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 IPSL-CM5A-LR_4.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_8.5 1000

900 Mulghat

800

700

600

500

400

Stream Flow CMS 300

200

100

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

60 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 27 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

6000 Chatara 5000

4000

3000

Stream Flow CMS 2000

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

Table 11 Dependable flow at various locations (CMS) for IPCC AR5 RCP 4,5 and RCP 8.5 scenarios (Baseline, and mid century)

Model Dependability Barhbise Pachuwar Khurkot Rabuwa Turkeghat Mulghat Chatara Ghat Bazar 10 566.5 684.8 1286.5 563.1 1388 768 4383 25 145.1 527.5 920.5 362.8 1042.2 579.3 3481.1 Observed 50 69.9 181.9 365.9 128.2 418.9 277.2 1319.5 75 32.3 66 123 46.1 127.1 112.9 530.8 90 21.3 40.2 69.1 21.4 60.2 68.2 276.3 10 628.6 794.9 1450 583.3 1418.5 813 4826.5 25 202.7 609.3 1036.2 384 1033.4 588.6 3698.7 CanESM2_4.5 50 109.2 232.4 412.2 136.1 459.4 268.8 1449.5 75 52.4 92 157.6 49.2 140.5 120.1 592.7 90 35 54.9 92.3 23.3 72.6 75.3 341.5 10 639.1 836.2 1611.5 654.5 1570 902.4 5515 25 220.2 640 1158.8 443.5 1206.7 665 4063.7 CanESM2_8.5 50 123.5 250.5 466 146.2 481.5 264.6 1507 75 57.5 99.4 165 49.6 142.5 108.8 617.2 90 37.5 59 94.4 23.7 70.9 65.9 331.4 10 611.3 761.9 1370 578.1 1437 803.9 4638.5 25 185.2 588.5 1012.4 369.2 1079 600.9 3636.9 CCSM4_4.5 50 94.3 225.9 417 140.7 472.1 301.8 1438.5 75 45.4 85.6 157.4 49 143.7 126.6 658.6 90 30.4 56.2 92.3 24.2 72.1 74.9 337.7 10 619.9 894.1 1646.5 611.4 1521 826.2 5126.5 CSIRO-Mk3-6- 25 193.2 649.3 1118.6 389.5 1076.7 593.2 3861.6 0_8.5 50 102.8 255.1 486.4 142.3 476.6 282.4 1526 75 48.1 89.6 163.1 46.8 145.7 114.4 611.1

61 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Model Dependability Barhbise Pachuwar Khurkot Rabuwa Turkeghat Mulghat Chatara Ghat Bazar 90 31.3 53.7 86.2 19 75.7 72.4 303.5 10 613.7 808.5 1432.5 587.9 1418 814.2 4630 25 192.7 599.8 997.8 367.8 1025.5 612.2 3593.5 GFDL- 50 97.3 237 453.4 150.4 527.2 316.7 1689.5 ESM2G_8.5 75 47.5 94.2 163.9 48.5 139 123.9 631.6 90 31 55.5 87.6 25.4 70.8 75.5 329.7 10 613.6 797.2 1515 585.4 1464 826.1 4919 25 189.9 608 1099.9 413.9 1059.7 620.4 3780.5 GISS-E2-R_4.5 50 97.1 228.1 441.2 134.9 450.3 292.6 1404 75 46.2 84 153.1 45 137.1 121.1 625.4 90 30.6 54.9 88 21.6 67.2 71.8 320.4 10 618.6 739.8 1305.5 540.4 1308.5 775 4356 25 188.5 571.4 934.2 337.1 931.9 548.8 3251.7 IPSL-CM5A- 50 95.4 204.7 367.9 128.6 419.1 266.4 1366.5 LR_4.5 75 47.8 80.1 141.2 43.4 125.1 114.5 566 90 31.5 51.3 83.5 21.1 65.1 74 308.9 10 631.1 831.8 1551 654.2 1474 853.5 5158.5 25 214.8 640.6 1121.7 413.6 1000.6 578.2 3927.8 IPSL-CM5A- 50 107.6 232 425.5 133.4 418 247.3 1402.5 LR_8.5 75 54.5 85.4 149 41.7 131.2 99.1 526.4 90 34.5 51.9 79.8 17.7 69.5 65.6 294

Figure 28 shows simulated 75% dependable flows at 7 locations. It can be seen that in most of the locations 75% dependable flow is higher in mid century period as compared to the baseline. Exception is found at Rabua Bazar (GISS-E2-R_4.5, IPSL-CM5A-LR_4.5, IPSL-CM5A-LR_8.5), Turkeghat (IPSL-CM5A-LR_4.5), Mulghat (CanESM2_8.5, IPSL-CM5A-LR_8.5) and at Chatara (IPSL-CM5A- LR_8.5).

Figure 28 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

Barhbise on Bhote Koshi 70 60 50 40 30 20 10 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

62 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 28 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

Pachuwar Ghat on Sun Koshi 120 100 80 60 40 20 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Khurkot on Sun Koshi 180 160 140 120 100 80 60 40 20 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

63 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 28 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

Rabuwa Bazar on Dudhkoshi 52 50 48 46 44 42 40 38 36 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Turkeghat on Arun 150 145 140 135 130 125 120 115 110 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

64 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 28 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for seven locations in Koshi River basin

Mulghat on Tamor 140 120 100 80 60 40 20 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Chatara on Saptkoshi 700 600 500 400 300 200 100 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

65 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Climate Change Impact on flow for Ganga Basin Assessment of dependable lean season flows along with their distribution in time is essential for planning and development of water supply schemes. The impact of the climate change on the water yield of the Ganga river system has been analyzed with respect to three levels of 50, 75 and 90% dependability. The 90% exceedance probability level is considered safe for determining assured water supply. Figure 29 shows the locations where flow dependability is assessed for both RCP 4.5 and RCP 8.5 scenario towards mid century.

Figure 29 Locations where Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for the Ganga River basin

Impact of climate change on the dependable flow at various locations (Figure 29) is given in Table 12. With the available climate change projections it can be inferred that the supply at these locations may gain considerably at lower dependable flows such as 10%, 20% and may also have marginal increase at higher dependability levels such as the 75% dependable flow level (which is the conventional dependability level for many water resources designs) for IPCC RCP 4.5 and RCP 8.5 scenarios for the mid century. This may be attributed to the additional lean season flow becoming available because of glacial melt and therefore should be a cause of concern in the long run.

66 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

1800 1600 Ramganga

1400 1200 1000 800 600 StreamFlow CMS 400 200 0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

7000 Upper Ganga 6000

5000

4000

3000

StreamFlow CMS 2000

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

67 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

25000 Gomti to Ghaghra 20000

15000

10000 StreamFlow CMS 5000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

40000

35000 Ghagra

30000

25000

20000

15000

StreamFlow CMS 10000

5000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

68 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

7000 Gandak 6000

5000

4000

3000

2000 StreamFlow CMS

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

900 Burhi Gandak 800 700

600 500 400 300

StreamFlow CMS 200 100 0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

69 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

6000 Koshi 5000

4000

3000

2000 StreamFlow CMS

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

6000 Chambal 5000

4000

3000

2000 StreamFlow CMS 1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

70 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

14000 Yamuna 12000

10000

8000

6000

4000 StreamFlow CMS 2000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

1200 Tons 1000

800

600

400 StreamFlow CMS 200

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

71 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

4000 Sone 3500

3000

2500

2000

1500

1000 StreamFlow CMS

500

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

450 400 Punpun 350

300 250 200 150

StreamFlow CMS 100 50 0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

72 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

600 Kiul 500

400

300

200 StreamFlow CMS 100

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

60000 Farakka 50000

40000

30000

20000

StreamFlow CMS 10000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

73 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 30 Simulated Dependable Flow (based on average monthly flows) for Observed climate and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at various locations in Ganga River basin

1200 Damodar 1000

800

600

400 StreamFlow CMS 200

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed (Aphrodite) CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_RCP8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

Table 12 Dependable flow at various locations on Ganga (CMS) for IPCC AR5 RCP 4,5 and RCP 8.5 scenarios (Baseline, and mid century)

Model Depend Ramganga Upper Gomti Gomti to Ghagra Gandak Burhi Koshi ability Ganga Ghaghra Gandak Observed 10 229.9 1845 428.8 8623 11940 3771.5 364.1 4074.5 (Aphrodite) 25 58.2 941.4 38.8 1661.1 3425.8 2468.8 71.4 2714.6 50 5.7 351.1 0 228.4 313.7 850.4 0 827.8 75 0 175.7 0 53 87.8 349.8 0 462.2 90 0 106.1 0 8.3 17.9 191.9 0 308.4 CanESM2_4 10 776 3538 757.5 12615 19500 4013.5 421.8 5062.5 .5 25 225.3 1470.5 115.9 2830.4 5944.2 2414.2 55.2 3033.8 50 27.6 538.5 0 415.3 529.2 956.1 0 1159.5 75 7.6 256.1 0 126 169.8 373.4 0 565.8 90 0 167.1 0 49.2 57.5 225.1 0 407.6 CanESM2_8 10 772.3 3368.5 626.8 11640 17845 4190.5 476.9 5413 .5 25 193.6 1417.9 75 2280.2 4975.4 2459.9 56.9 3452.1 50 27.5 571.6 0 438.9 546.7 951.8 0.4 1258.5 75 5.1 286.8 0 156.6 194.9 376.1 0 617.8 90 0 193 0 66.3 66.9 217.2 0 438.2 CCSM4_4.5 10 684.1 3122.5 507.1 11525 16395 4025 495.8 4808 25 218.2 1297.5 65.5 2210.4 4959.6 2713.4 128.9 3135.6

74 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Model Depend Ramganga Upper Gomti Gomti to Ghagra Gandak Burhi Koshi ability Ganga Ghaghra Gandak 50 27.5 462.9 0 349 532.6 929.9 1 1004.5 75 5.8 240.9 0 125 166.3 412 0 585.8 90 0 153.8 0 42.3 65.3 235.7 0 391.9 CSIRO- 10 1710 5771.5 1321 21890 35090 5969.5 809.9 5687 Mk3-6- 25 455.6 2013.1 314.6 4728.9 11480.8 3933.8 238.7 3495.6 0_8.5 50 48.8 525.2 0 447.8 828.1 1154 15.1 1123.5 75 12.4 256.9 0 133.1 246.5 458.3 1.2 580.1 90 0.5 165 0 36.1 68.3 267 0 397.9 GFDL- 10 371 2097 341.3 7244.5 12015 4017.5 436.3 4931.5 ESM2G_RC 25 74 1062.8 24 1468 3068.6 2697 74.3 3011.8 P8.5 50 16.1 409.4 0 279.6 389.9 995.7 3.6 1120.5 75 0 194.9 0 79.3 117.5 374.2 0 573.6 90 0 125.3 0 16.5 24.7 224.3 0 389.4 GISS-E2- 10 949.3 3984.5 1383 15955 23320 4705.5 803 4931 R_4.5 25 272.4 1536.6 297.5 3661.6 6878.5 3184.9 211 3281.5 50 32.2 473.9 0.1 365.1 490.4 1015 18 1027 75 7.4 234.2 0 122.4 160.6 406.2 2 570.4 90 0 162.9 0 34.2 57.5 237.9 0 384.2 IPSL-CM5A- 10 562.5 2503 609.2 10700 16255 3910.5 478.4 4780 LR_4.5 25 154 1156 75.3 2291.5 4836 2628.1 108 2801.6 50 22.4 383 0 268.9 411.8 916.9 1.1 970.3 75 0.6 196.1 0 69.1 106.9 367.2 0 511.3 90 0 121.3 0 6.8 21.5 217.3 0 359 IPSL-CM5A- 10 1216.5 3863 626.9 14510 21940 4399 666.6 5484.5 LR_8.5 25 219.7 1414.3 103.7 2909 6518.9 2748.6 141.4 3308.2 50 27.6 542.9 0 461.7 555.4 926.6 5 1149 75 6.4 264 0 148.5 179.6 359.9 0 569.6 90 0 185.4 0 53.8 60.5 218.1 0 397.9

Model Dependabilit Chambal Yamuna Tons Sone Punpun Kiul Farakka Damodar y Observed 10 928.3 4043.5 654.2 1388.5 147.7 194.1 25005 631.3 (Aphrodite) 25 43.5 427.9 47.1 374.1 22 29.7 11892.4 258.1 50 0 0 0 263.4 0 0 2095 0 75 0 0 0 219.4 0 0 996.5 0 90 0 0 0 193.7 0 0 696.2 0 CanESM2_4 10 1680.5 6002.5 844.6 1835 181.2 248.9 34785 678.9 .5 25 91.4 730.8 72.6 479.8 16.7 28.5 15588 174.2 50 0 0 0 262.6 0 0 2735 0

75 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Model Dependabilit Chambal Yamuna Tons Sone Punpun Kiul Farakka Damodar y 75 0 0 0 218.5 0 0 1218 0 90 0 0 0 192.3 0 0 883.9 0 CanESM2_8 10 1503 5499 671.2 1562.5 179.3 265.8 33910 654.4 .5 25 70.2 573.2 68 434.3 14.5 28.8 13948.8 173.6 50 0 0 0 261.4 0 0 2895.5 0 75 0 0 0 222.3 0 0 1255.7 0 90 0 0 0 191.9 0 0 947.1 0 CCSM4_4.5 10 1404 5402 621.9 1614.5 205.9 296 31375 882.3 25 57.2 545.1 94.6 443.3 31.5 51 14691.2 246.6 50 0 0 0 263.1 0 0 2480 0 75 0 0 0 213.3 0 0 1271.6 0 90 0 0 0 193.1 0 0 939.2 0 CSIRO- 10 4904 12040 898 3430.5 379.2 483.5 55955 1126 Mk3-6- 25 593 1934.2 155.9 997.9 67 87.1 23499.2 357.6 0_8.5 50 0 0.1 0 263.1 0 0 3285 0 75 0 0 0 233.7 0 0 1374.1 0 90 0 0 0 193.1 0 0 952.3 0 GFDL- 10 565.9 2779.5 486.5 763.7 141.7 214.9 25205 670.9 ESM2G_RC 25 2.4 201.6 23.6 306.6 11.3 19.6 11356 107.1 P8.5 50 0 0 0 260.6 0 0 2707 0 75 0 0 0 204.5 0 0 1182.5 0 90 0 0 0 192.6 0 0 871.8 0 GISS-E2- 10 1605.5 6991 1013.5 3510 413.5 475.3 42285 812.2 R_4.5 25 73.7 734.1 225.7 943.8 110 113.7 18400 275.2 50 0 0 0 264.6 0 0 2748 0.1 75 0 0 0 246.3 0 0 1259.2 0 90 0 0 0 193.7 0 0 912.1 0 IPSL-CM5A- 10 1661 4944.5 651.9 1602.5 250.3 392.7 31865 1068.5 LR_4.5 25 81 622.2 97.7 475.6 38 66.9 13965.6 296.3 50 0 0 0 262.9 0 0 2393.5 0 75 0 0 0 229.1 0 0 1095 0 90 0 0 0 193.2 0 0 787.2 0 IPSL-CM5A- 10 3034 7989.5 758.2 2793 288 399.9 41730 1033 LR_8.5 25 241.6 972.4 114.7 726.8 42.8 78.4 17399.6 266 50 0 0 0 262.2 0 0 2775 0 75 0 0 0 245 0 0 1171.7 0 90 0 0 0 192.7 0 0 852.1 0

76 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 31 shows simulated 75% dependable flows at 7 locations. It can be seen that in most of the locations 75% dependable flow is higher in mid century period as compared to the baseline which reflects some additional contribution coming from the glacial melt during the mid century period. Therefore this cannot be a good situation for the basin in the long run since it is not sustainable.

Figure 31 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for various locations in Ganga River basin

Ramganga

14 12 10 8 6 4 2 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Upper Ganga

350 300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

77 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 31 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for various locations in Ganga River basin

Gomti to Ghaghra

180 160 140 120 100 80 60 40 20 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Ghagra

300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

78 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 31 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for various locations in Ganga River basin

Gandak

500 450 400 350 300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Koshi

700 600 500 400 300 200 100 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

79 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 31 Simulated 75% Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for various locations in Ganga River basin

Sone

300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Farakka

1600 1400 1200 1000 800 600 400 200 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

80 Chapter 4 - Climate Change Scenarios using - Delta Change Method weather data | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 5

Future Demand Scenarios and Impact of Climate Change

81 Chapter 5 | IIT Delhi

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 5 - Future Demand Scenarios and Impact of Climate Change

Future Water Demand Scenario Future planned projects on Koshi basin in Nepal has been implemented with operation. The projects implemented is given in Table 13.

Table 13 Details of the Future development Projects

Status Type Name Country River Latitude Longitude Gross Ma Submerged Hydro Storage height area (sq power Capacity (m) km) capacity (MCM) (MW) Planned Dam Sun Nepal Sun 27.16 86.43 1500 120 31 0 Kosi Kosi Dam I Planned Dam Sun Nepal Sun 27.1 86.49 4400 166 0 0 Kosi Kosi Dam II Planned Dam Tamur Nepal Tamur 26.96 87.43 1890 150 32 250 Dam Planned Dam Sapt Nepal Kosi 26.84 87.14 13450 134.5 195 3489 Kosi

Three demand scenarios are used as given in Table 10.

Table 14 Scenarios used in the study

Scenario Years Representing Major Water Operation Diversions Infrastructure Dams in Nepal 2020-30 Sun Koshi dam, Hydropower, flood Additional Scenario D5 Tamor dam and mgmt, and lowflow diversions 5,10,15 Scenario D10 Sapta Koshi Dam augmentation BCM/yr split in rabi, Scenario D15 summer & kharif

The above scenario run using statistical down scaled climate scenario data provided by ICIMOD (Future Water). 12 grids with monthly delta change data for the future (2021-2050) relative to a reference period (1961-1990) to reflect the change in temperature (K) and precipitation (%) over 60 years was provided. Transient time series was generated using the monthly delta change grids using the observed rainfall (station as well as gridded Aphrodite) and temperature34.

The Hydrologic Simulation with Climate Change Scenarios The SWAT hydrological model has been run using climate scenarios for near term (MC, 2006 – 2051) without changing the land use. The simulations under climate change are run for 2 RCP scenarios

34 Manual to generate downscaled climate change scenarios for the downstream basins of the HICAP project, 2012, W.W. Immerzeel 82 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

(RCP 4.5 and RCP 8.5). Three water demand scenarios of 5 BCM/year, 10 BCM/year and 15 BCM/years have been included in these runs.

The outputs of these scenarios have been analyzed to evaluate the possible impacts on the runoff, snowmelt, baseflow, soil moisture, ground water recharge and actual evapotranspiration (expressed as change between the baseline and future periods).

Flow Duration Curve and Impact on dependable flow for Koshi Basin Assessment of dependable lean season flow along with their distribution in time is essential for planning and development of water supply schemes. The impact of the climate change on the dependability of the water yield of the Koshi river system has been analyzed with respect to three levels of 50, 75 and 90% dependability. The 90% exceedance probability level is considered safe for determining assured water supply. Figure 24 shows the locations where flow dependability is assessed for both RCP 4.5 and RCP 8.5 scenario towards mid century. Since the planned reservoirs are on Sun Koshi and Tamour rivers, the impact of three demand scenario is assessed at Mulghat and Chatara.

Figure 32 Locations of Planned Reservoirs and Locations where where Dependable Flow is assessed for BL and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5) for the Koshi River basin

Impact of climate change on flow at Mulghat and Chatara is shown in Figure 24 is given in Table 9 for various dependable levels. With the available climate change projections it can be inferred that the flow availability at these locations may not have much reduction at the 75% dependable flow level

83 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]] for IPCC RCP 4.5 and RCP 8.5 scenarios. However marginal decrease at 75% dependability levels is projected for RCP 8.5 scenario for 2 climate models.

Figure 33 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at Mulghat in Koshi River basin Demand Scenarios- Mulghat

1200

Mulghat 1000

Demand Scenario 5 BCM/year 800

600

400 StreamFlow CMS

200

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%)

Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

1000 Mulghat 800 Demand Scenario 10 BCM/year

600

400

200 StreamFlow CMS

0 0 10 20 30 40 50 60 70 80 90 100

-200 Dependability (%)

Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

84 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 33 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at Mulghat in Koshi River basin

1000 900 Mulghat 800

700 Demand Scenario 15 BCM/year 600 500 400 300 StreamFlow CMS 200 100 0 0 10 20 30 40 50 60 70 80 90 100 -100 Dependability (%)

Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

Figure 34 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at Charata in Koshi River basin Demand Scenarios Chatara

7000 Chatara 6000

5000 Demand Scenario 5 BCM/year 4000

3000

StreamFlow CMS 2000

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

85 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 34 Simulated Dependable Flow for Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) at Charata in Koshi River basin

7000

6000 Chatara

5000 Demand Scenario 10 BCM/year 4000

3000

StreamFlow CMS 2000

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

7000

Chatara 6000

5000 Demand Scenario 15 BCM/year 4000

3000

StreamFlow CMS 2000

1000

0 0 10 20 30 40 50 60 70 80 90 100 Dependability (%) Observed CanESM2_4.5 CanESM2_8.5 CCSM4_4.5 CSIRO-Mk3-6-0_8.5 GFDL-ESM2G_8.5 GISS-E2-R_4.5 IPSL-CM5A-LR_4.5 IPSL-CM5A-LR_8.5

86 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Table 15 Dependable flow at Mulghat and Chatara locations (CMS) for IPCC AR5 RCP 4,5 and RCP 8.5 scenarios (Baseline, and mid century) for 3 demand scenarios

5 BCM/year 10 BCM/year 15 BCM/year Model Dependability Mulghat Chatara Mulghat Chatara Mulghat Chatara Observed 10 735.9 4766.5 698.1 4660.5 652.1 4538 25 528.8 3499.5 492.4 3365.8 436 3105.5 50 189.8 1182.5 125.8 977.3 32.5 830.4 75 25.8 328.3 2 225.5 0.9 180.4 90 1.5 99.8 0.2 70.5 0.2 64 CanESM2_4.5 10 828.2 5234.5 791.2 5128 754.8 5011.5 25 590.3 3595 554.2 3487.4 512.3 3377.8 50 215.2 1278 158.3 1103 70.5 901 75 33.3 355.3 2.8 241.4 1.1 194 90 3.4 115 0.3 73.5 0.2 65.9 CanESM2_8.5 10 952.9 5964.5 917.7 5858.5 882.2 5750 25 691.7 4223.4 651.8 4116.7 605.4 3927.5 50 224 1354 170.1 1165 80.4 941.5 75 38.2 339.8 3.1 223.9 1.1 193.5 90 4.3 118.6 0.3 76.2 0.2 63.3 CCSM4_4.5 10 776.7 5100.5 738.2 4993 703.1 4886 25 564.3 3671.2 528.4 3577.3 488.7 3432.7 50 211.6 1313.5 159.7 1110 64.3 929.2 75 32.9 399.1 4.2 299.4 1.2 216.8 90 4.5 130.5 0.3 79.8 0.2 76.7 CSIRO-Mk3-6- 10 820.9 5925 778.2 5819 742.6 5706.5 0_8.5 25 583.6 4241.9 543 4130.8 471 3863.2 50 194.7 1414.5 120.4 1194 31.8 1019 75 23.6 366.5 1.2 271.3 0.7 230.2 90 1 115.9 0.2 73.8 0.1 63.4 GFDL- 10 784.7 5174.5 746.9 5069 711.9 4963.5 ESM2G_8.5 25 583 3718.1 542.8 3609.3 500.7 3470.9 50 224.4 1660.5 169.8 1451.5 106.3 1223.5 75 32.7 372.6 3.9 269.2 1.1 189.4 90 4.1 113.2 0.3 70.3 0.2 67.2 GISS-E2-R_4.5 10 808.8 5488.5 773.4 5379.5 736.7 5233.5 25 579.9 3988.3 542.9 3882.8 503.5 3777.4 50 193.5 1333.5 144 1163.5 53.8 945.9 75 30.3 370.2 2.5 274.4 0.9 208.5 90 2.7 122 0.3 75.6 0.2 72 IPSL-CM5A- 10 761.9 4804.5 726.2 4699 726.2 4699 LR_4.5 25 535.2 3331 495.8 3211.7 495.8 3211.7

87 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

5 BCM/year 10 BCM/year 15 BCM/year Model Dependability Mulghat Chatara Mulghat Chatara Mulghat Chatara 50 203.4 1253.5 153.2 1043.5 153.2 1043.5 75 29.7 315.4 2.2 222.7 2.2 222.7 90 2.1 107 0.3 65.5 0.3 65.5 IPSL-CM5A- 10 849 5578.5 813.2 5467 813.2 5467 LR_8.5 25 607.1 3821.4 569.4 3695.4 569.4 3695.4 50 204 1248 136 1043 136 1043 75 26.5 291.1 1.4 196.4 1.4 196.4 90 1.4 95.1 0.2 65.7 0.2 65.7

Figure 28 shows simulated 75% dependable flows at Mulghat and at Chatara for all three demand scenarios. It can be inferred from the figure that the demand scenario is not sustainable at Mulghat (Tamour dam) beyond 5 BCM/year. Reduction of about 60% in stream flow is projected when 5 BCM/year demand is imposed on the business as usual scenario. No negative impact of climate change at this location is projected for all climate models except CSIRO-Mk3-6-0 for RCP 8.5.

Similarly at Chatara (after, Sun Koshi 2 dams, Tamour dam and Sapt Koshi dam) reduction of about 21%, 45% and 56% in stream flow is projected with 5, 10 and 15 BCM/year respective demand is imposed on the business as usual scenario. No negative impact of climate change at this location is projected for all climate models except IPSL-CM5A-LR for RCP 8.5

Figure 35 Simulated 75% Dependable Flow for Observed bAU, Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for Mulghat and Chatara in Koshi River basin with 3 demand scenarios 75% Dependable flow for 3 Demand Scenarios - Mulghat

Mulghat on Tamor 80 70 Demand Scenario 5 BCM/year 60 50 40 30 20 10 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

88 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 35 Simulated 75% Dependable Flow for Observed bAU, Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for Mulghat and Chatara in Koshi River basin with 3 demand scenarios

Mulghat on Tamor 80 Demand Scenario 10 BCM/year 70 60 50 40 30 20 10 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Mulghat on Tamor 80 Demand Scenario 10 BCM/year 70 60 50 40 30 20 10 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

75% Dependable flow for 3 Demand Scenarios - Chatara

89 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 35 Simulated 75% Dependable Flow for Observed bAU, Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for Mulghat and Chatara in Koshi River basin with 3 demand scenarios

Chatara on Saptkoshi 450 Demand Scenario 5 BCM/year 400 350 300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

Chatara on Saptkoshi 450 400 Demand Scenario 10 BCM/year 350 300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

90 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Figure 35 Simulated 75% Dependable Flow for Observed bAU, Observed and MC climate scenarios (IPCC AR5 RCP 4.5 and RCP 8.5, statistical downscaling) for Mulghat and Chatara in Koshi River basin with 3 demand scenarios

Chatara on Saptkoshi 450 400 Demand Scenario 15 BCM/year 350 300 250 200 150 100 50 0 75% Dependable Stream Flow Flow CMS Stream Dependable 75%

91 Chapter 5 - Future Demand Scenarios and Impact of Climate Change | IIT Delhi.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 5

Summary and Conclusions

92 Chapter 5 | INRM Consultants.

DRAFT FINAL REPORT [SWAT Hydrological Modelling]]

Chapter 5 - Summary and Conclusions

The present study is aimed at water availability assessment of the Koshi River Basin and Ganga river basin using the hydrological simulation methodology. The well known distributed hydrological model SWAT has been deployed for the purpose. Scenarios have been generated for assessing implications on quantity of water on account of climate change scenarios.

The data available from international sources on terrain and landuse, soil have been used to set up the base SWAT model for the Koshi river basin and Ganga river basin . The ArcSWAT GIS interface has been used to pre-process the spatial data for the Koshi and Ganga river basin river system. The basin has been sub-divided into 47 sub-basins to adequately account for the spatial variability of various inputs and outputs.

The daily rainfall data received from ICIMOD for 17 stations (1970-2010) and APHRODITE35 gridded data at a resolution of 0.25° X 0.25° latitude by longitude grid points (1971–2006) has been used. The daily maximum and minimum temperature data received from ICIMOD for 15 stations (1970-2010) and NCEP36 temperature gridded data at a resolution of 0.5° X 0.5° latitude by longitude grid points (1979– 2010) has been used. Model has been calibrated at seven locations. The validation was found to be satisfactory, thereby instilling confidence in the simulation process that is going to be the basis of development and future scenarios.

While looking at the impact of climate change the baseline scenario of IPCC AR5 RCP 4.5 and RCP 8.5 scenario from BCCR RCM outputs was run first followed with the mid century scenarios. Statistical downscaled 8 climate model outs for RCP 4.5 and RCP 8.5 scenarios were also run.

The model has been run using climate scenarios for near term (MC, 2006 – 2051) without changing the land use. The climate change simulations are run for 2 RCP scenarios (RCP 4.5 and RCP 8.5), for 2 time periods; Baseline and Mid Century.

The distribution of the major annual water balance components as percentage change in response to the change in precipitation was analysed for each of the simulations. Impact for extreme events of drought frequency and flood magnitude in future period was analysed. Subsequently projected impact on reservoirs in terms of dependable flows was carried out.

Three demand scenarios on 4 future reservoir projects in Koshi basin was implemented and compared its impact on the downstream locations. Climate change imnpact combined with future demand scenario was implemented.

35 http://www.chikyu.ac.jp/precip/products/index.html 36 The National Centers for Environmental Prediction (NCEP) Climate Forecast System Reanalysis (CFSR), http://globalweather.tamu.edu/ 93 Chapter 5 - Summary and Conclusions | INRM Consultants