See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/309589915

Modeling the Agricultural Water–Energy–Food Nexus in the Indus River Basin,

Pakistan

Article  in  Journal of Water Resources Planning and Management · August 2016

DOI: 10.1061/(ASCE)WR.1943-5452.0000710

CITATIONS READS
62 1,620

4 authors:

Yi-Chen Ethan Yang Claudia Ringler

Lehigh University International Food Policy Research Institute
82 PUBLICATIONS   1,837 CITATIONS    289 PUBLICATIONS   18,115 CITATIONS   

SEE PROFILE SEE PROFILE

Casey Brown Md Alam Hossain Mondal

University of Massachusetts Amherst International Food Policy Research Institute
172 PUBLICATIONS   5,907 CITATIONS    51 PUBLICATIONS   2,279 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Quantifying Impact of Water Variability on Global Economic Growth View project

Climate-Informed Estimation of Hydrologic Extremes for Robust Adaptation to Non-Stationary Climate View project

All content following this page was uploaded by Yi-Chen Ethan Yang on 26 October 2017.

The user has requested enhancement of the downloaded file.

Manuscript Click here to download Manuscript Indus-water-energy-text-
0426.docx

1 Modeling the Agricultural Water-Energy-Food Nexus in the Indus River Basin of Pakistan

3 Y. C. Ethan Yang1* Claudia Ringler2, Casey Brown3 and Md. Alam Hossain Mondal4

1*
5 Corresponding Author, Research Assistant Professor, Department of Civil and Environmental

6 Engineering, University of Massachusetts – Amherst, 130 Natural Resources Rd Amherst MA

7 01002 USA. [email protected], PH: 413-577-3232, FAX: 413-545-2202

2
8 Deputy Division Director, International Food Policy Research Institute, 2033 K St, NW

9 Washington, DC 20006 USA

3
10 Associate Professor, Department of Civil and Environmental Engineering, University of

11 Massachusetts – Amherst, 130 Natural Resources Rd Amherst MA 01002 USA.

4
12 Research Fellow, International Food Policy Research Institute, 2033 K St, NW Washington, DC

13 20006 USA

14

15 Abstract

16 Nexus thinking is critical to jointly address growing water, energy and food security

17 challenges. This paper evaluates the water, energy and food nexus (WEFN) in the Indus River of

18 Pakistan using the Indus Basin Model Revised – Multi Year, a hydro-agro-economic model

19 extended with an agricultural energy use module. We model impacts of a range of climate change

20 scenarios on the WEFN in the Indus Basin and then assess the potential of different alternative

21 water allocation mechanisms and water infrastructure developments to address growing water,

22 energy and food security concerns in the country. We find growing water and energy use under

23 hotter and wetter climate conditions. While more flexible surface water allocation policies can

1
 1 mitigate negative climate change impacts on agricultural water and energy use allowing for larger

2 crop and hydropower production, such policies might also increase the inter-annual variability of

3 resource use. Moreover, a more flexible surface water allocation policy would increase surface

4 water use in the basin, while groundwater and energy use would be lower. Study results can inform

5 the WEFN in areas with similar hydro-climatic environments, such as California and Central Asia.

6 Further integration of a groundwater model and an energy market model and explicitly addressing

7 changes in food and energy demand as a result of demographic dynamics are three areas for future

8 study.

9 Keywords: systems analysis, climate change, water infrastructure, water allocation

10
11

2
 1 1. Introduction

2 Growing water, energy and food needs are concerns in both developed and developing

3 countries. With growing population and economic development, 50% more food, 40% more

4 energy and 30% more water will be needed to satisfy global demands by 2030 (Winston, 2013).

5 Solutions with “nexus thinking” are critical and necessary to address these challenges

6 simultaneously and in a systems context (Biggs et al., 2015). Nexus thinking aims to link water,

7 energy and food and provide tools to increase resource use efficiency. It ensures policy coherence

8 and coordination across sectors and stakeholders to build synergies and generate co-benefits (Yang

9 et al., 2016). The objective of this paper is to develop and apply a modeling approach that explicitly

10 simulates the water-energy-food nexus (WEFN), highlights the impacts of climate change on the

11 nexus and suggests a way forward with nexus thinking.

12 Modeling approach have been used for WEFN analysis in large river basins in Asia such

13 as the Mekong (Räsänen et al., 2014) and the Brahmaputra (Yang et al., 2016). This article uses

14 the Indus River of Pakistan as a demonstration. The Indus River is critical to Pakistan’s water

15 supply, food production and energy generation. Irrigation is essential for both food production and

16 agricultural employment in Pakistan; about 90% of staple crops and 100% of cash crops produced

17 in the country rely on Indus basin waters (Ringler and Anwar, 2013). The state of food security in

18 Pakistan has deteriorated significantly since 2003, with almost half of the population in Pakistan

19 (48.6 percent) without access to sufficient food (Bhutta, 2011). Previous studies (e.g., Briscoe and

20 Qamar, 2006; Yu et al., 2013) have evaluated groundwater development and the role of reservoirs

21 for the Indus River Irrigation System as well as the impact of climate change on crop production

22 but without direct linkages with the energy sector.

3
 1 Pakistan has faced serious energy deficits since 2005 as a result of growing domestic and

2 industrial demands, in addition to agricultural uses, combined with poor efficiency of the

3 generation system. The average energy deficit in 2011 was about 4,500 Mega-watt (MW) (Siddiqi

4 et al, 2012) and reached 6,000 MW during the summer period (WAPAD, 2011a). Most previous

5 studies on the energy sector have focused on traditional, energy-centric economic analyses with

6 limited consideration of water and food linkages. For example, IRG (2011) built an integrated

7 energy model, Pak-IEM, that encompasses the entire energy system of Pakistan and was designed

8 to assess the impacts of various options and strategies for meeting the country's future energy needs

9 in an optimal (minimum cost) manner. However, hydropower, as the primary future energy source

10 targeted by the government (Izhar-ul-Haq and Munir, 2009), is simplified in Pak-IEM. As such

11 there are directly linkages with water supply and food. Tate and Farquharson (2000) evaluated

12 sediment accumulation in Mangla and Tarbela and its effect on irrigation and hydropower using

13 modeling and data analysis. Ahmad (2009) and Khan and Tingsanchali (2009) used an

14 optimization model to assess how reservoir operation (Mangla and Tarbela) influences agricultural

15 and hydropower profits. Hussain (1993) and WAPDA (2011b) discussed the expansion or

16 modification of the original reservoir design to increase hydropower generation. Currently, total

17 installed hydropower capacity on the Pakistan Indus River is about 6,599 MW, accounting for 34%

18 of national energy supply (Mirza et al., 2008). In 2014, the completion of the Mangla Dam Raising

19 Project added 120 MW of installed capacity into the system. However, Amir (2005) suggested that

20 the water economy of the Indus Basin might change if dams are operated for hydropower rather

21 than irrigation as they are now.

22 In 2014, the country’s Planning Commission released the “Vision 2025” that states that

23 “Pakistan Vision 2025 recognizes that sufficient, reliable, clean and cost-effective availability of

4
 1 energy, water and food – for now and the future –is indispensable in ensuring sustainable

2 economic growth and development. These key sectors have suffered historically from severe

3 failings of integrated policy and execution” (Planning Commission, 2014, p. 59). To achieve this

4 vision, a systems analysis with nexus thinking approach to quantitatively evaluate water use for

5 crop production and hydropower generation is critical to support decision making (Scott et al.,

6 2011).

7 In this paper, we define WEFN from two aspects: 1) water/energy uses for agricultural

8 activities and 2) crop production and hydropower generation at basin-wide and provincial scale. A

9 hydro-agro-economic model: The Indus Basin Model Revised – Multi Year (IBMR-MY) with a

10 post-calculated energy use module is applied to evaluate the WEFN at basin-wide (Indus) and

11 provincial scale in Pakistan under different climate change, investment and water allocation

12 scenarios. Different temperature and precipitation combinations are tested following the concept

13 of decision scaling (Brown et al., 2012). New water infrastructures (e.g. the raised Mangla Dam

14 and planned Diamer-Basha Dam) and different surface water allocation policies (e.g. historical

15 and interprovincial Indus Water Accord) are also considered.

16 The modeling results demonstrate an in-depth assessment of how different natural and

17 socioeconomic scenarios affect the WEFN such as hotter climate, and flexible water allocation

18 policy results in tradeoff between surface water and energy uses and can also mitigate climate

19 change impacts. We will briefly introduce the model used in this paper in Section 2; describe the

20 baseline and tested scenarios in Section 3; demonstrate modeling results with an in-depth

21 assessment of how WEFN changes in Section 4; discuss the potential knowledge transfer to other

22 places in Section 5 and summarize entry points to improve water, energy and food security in a

23 basin context.

5
 1 2. The Indus Basin Model Revised – Multi Year

2 The Indus Basin Model Revised (IBMR) has been developed and updated by the Pakistan

3 government and the World Bank (Duloy and O’Mara, 1984; Yu et al., 2013; Yang et al., 2013;

4 NESPAK, 2013). It is a hydro-agro-economic model that covers four provinces: Khyber

5 Pakhtunkhwa, Punjab, Sindh and Balochistan and uses 12 agro-climatic zones (ACZs, Figure 1)

6 as basic spatial units. A node-link system is used to represent the river (supply node)-canal

7 (demand node) network, and this node-link system provides surface water to each ACZ.

8 Agricultural production and consumption (food commodity market with non-linear supply and

9 demand curve) is simulated at the ACZ level. Yang et al. (2014) expanded the IBMR into a multi-

10 year version and developed a preliminary groundwater module to address the underground water

11 balance at the ACZ level. The objective function maximizes the net economic benefit (from crop

12 production and hydropower generation) of water uses in the basin. The input data include: 1)

13 agronomic data, 2) economic data, 3) a resources inventory and 4) information on irrigation

14 systems. Primary modeling outputs include: net economic profit, crop and hydropower production

15 and surface water and groundwater usage at the provincial and basin-wide levels. The detailed

16 modeling framework is described in Yu et al. (2013), Yang et al. (2013) and NESPAK (2013).

17 To explicitly quantify the relationship between energy, water and food production, a

18 module to calculate agricultural energy uses (i.e. energy use for groundwater pumping and tractors)

19 is incorporated into the modeling structure. Domestic and industrial (D&I) water use from nine

20 major cities in Pakistan are included in the model with higher priority. However, overall water use

21 for D&I is relatively small and does not affect hydropower and agriculture water uses significantly.

22 The current model considers four major hydropower dams: Mangla, Tarbela, Chashma and Ghazi-

6
 1 Barotha. Specific characteristics, such as live capacity and installed capacity are shown in Table

2 S1 and their locations are shown in Figure 1.

3 The model is set up to run for a multi-year period using a monthly time step under a year-

4 by-year optimization scheme. The reservoir storage and groundwater depth at the end of each year

5 are used as initial conditions for the following year. Historical monthly inflows from 1961 to 2010

6 are used to set up the baseline condition. The model does not explicitly simulate demographic

7 dynamics and their effect on total food and energy demand to avoid further model complexity.

8 However, these dynamics could be assessed in a follow-up study.

9 3. Baseline and Scenarios

10 3.1. Model diagnosis with observed data

11 We assess model performance by recreating historical crop production and hydropower

12 generation. We source crop price information from the government report “Agricultural Statistics

13 of Pakistan 2008-2009” (MINFA, 2010) and electricity prices from the National Electric Power

14 Regulatory Authority (NEPRA, 2013). Tariffs range from 6 to 20 Rs/kWh. We use 10 Rs/kWh as

15 an average value in our model. Fifty years of streamflow data from the Indus main stem and major

16 tributaries are used as inputs for the baseline analysis. Other inputs such as crop water needs,

17 fertilizer and tractor use, crop price elasticities, available land area, canal capacity and pumping

18 well capacity, etc. are adapted from a previous version of the model (Yu et al, 2013).

19 Figure S1 in the supplemental materials compares modeling results with observed data.

20 Long-term average crop production in Punjab and Sindh (Figure S1a and b) and hydropower

21 generation of two major reservoirs, Mangla and Tarbela, and the system total (Figure S1c) are

22 compared. Observed crop production data are from MINFA (2010). Hydropower data for

23 individual reservoirs are collected from Ahmad (2009) and system total data are from Energy

7
 1 Information Administration (EIA, 2008). These results demonstrate that the model can capture the

2 historical condition of major crop production and hydropower generation patterns reasonably.

3 3.2. Scenario settings

4 3.2.1. Different water allocation policies

5 The first set of scenarios for WEFN analysis is related to alternative water allocation

6 policies. Historically, the warabandi system designated farmers to withdraw water at certain

7 days/hours and theoretically the risk of variability of flows was randomly shared. In this study, the

8 average 1991 to 2000 monthly canal diversions are used to represent this historical water allocation

9 scheme (“Historical”). A 20% deviation from these average monthly values is allowed in the model

10 to reflect changes in water withdrawals as a result of changes in inflow levels.

11 The 1991 interprovincial Indus Water Apportionment Accord between the four provinces

12 determines rules for water sharing in the basin. The fundamental concept of this Accord is to allow

13 intra-provincial freedom of canal allocations. However, this policy was not strictly enforced until

14 2003. The three-tiered water sharing mechanism between the four provinces described in Irshad

15 (2011) is used as the second water allocation policy to mimic the concept of the 1991 Accord that

16 has been enforced by the Indus River System Authority after 2003 (“IRSA”).

17 A third water allocation policy we test in this paper optimizes basin-wide water allocation

18 based on the marginal value of water (“Basin”) and thus disregards the constraints of the 1991

19 Accord. While physical constraints of diversion capacity on canal flows still apply, this water

20 allocation assumes everyone in the system follows a “perfect water market” with no transaction

21 costs. Informal water markets do exist in Pakistan but are currently limited to a micro scale and

22 mostly focus on groundwater but trading of canal turns is also observed (Qureshi et al., 2003;

23 Mekonnen et al., 2016). Therefore, while this policy does not replicate what is currently possible

8
 1 in practice, it does provide insights as to the full potential of economics-driven water allocation in

2 the basin and serves as an upper limit of the system.

3 3.2.2. New water infrastructures

4 We consider two new water infrastructure developments. The first one is the Mangla Dam

5 Raising Project completed in early 2014. According to WAPDA (2012), Mangla was raised by 30

6 feet (9.14 meters) with an additional 2.88 MAF, million acre-feet (or 3.55 BCM, billion cubic

7 meter) of live storage and 120 MW of installed capacity. For our purposes, we call this the “New

8 Mangla” scenario. The second (planned) investment is Diamer-Basha Dam. This dam is proposed

9 to be located on the Indus River 315 km upstream of Tarbela Dam. It has 6.39 MAF (7.88 BCM)

10 of live storage and installed capacity of 4,500 MW. Detailed operational data used in the paper

11 were obtained from Luna and Jabbar (2011). The joint infrastructure scenario is called “New

12 Mangla plus Basha.” Together, these new water infrastructures will add 9.28 MAF (11.43 BCM)

13 of storage into the system and increase installed capacity by 4,620 MW. The features of these new

14 dams are given in Table S1 and their locations are marked in Figure 1.

15 3.2.3. Climate change

16 We apply the decision scaling method (Brown et al., 2012) to evaluate the “response” of

17 the system under different temperature and precipitation conditions. Yu et al. (2013) used a glacier

18 mass balance model to evaluate the first order effect of temperature increases and precipitation

19 changes on streamflow. Thirty-six possible temperature (range from 0.5o C to 4.5o C) and

20 precipitation (range from -20% to 20%) change combinations are tested for their effects on

21 streamflow. These streamflow variations are directly used in this paper. Corresponding changes in

22 crop water requirements are also incorporated in the analysis. Climate change projections from the

23 Intergovernmental Panel on Climate Change (IPCC) ensemble of GCMs, the Third and Fifth Phase

24 of the Coupled Model Intercomparison Project (CMIP3 and CMIP 5), were consulted to inform

9
 1 the likelihood of changes in precipitation and temperature. We summarize the ensemble mean of

2 temperature and precipitation changes for two CMIP 3 scenarios and four CMIP 5 scenarios in the

3 near (2011-2040), middle (2041-2070) and far (2071-2099) future in Table S2.

4 4. Results

5 To quantify the influence of climate change, infrastructure development and water

6 allocation on the WEFN, we present results of 1) basin-wide water/energy use, 2) basin-wide crop

7 production and hydropower generation, 3) provincial level water/energy use and 4) provincial level

8 crop production under different scenarios. Results of climate change impacts on system

9 performance and historical allocation are shown first (Table 1 and Table 3), followed by the

10 combined effects of all scenarios (Figure 2 to Figure 5).

11 4.1. Basin-wide results

12 4.1.1. Impact of climate change on resource use

13 Table 1 presents annul mean and inter-annual variability changes of surface water use for

14 crops (SWC) and groundwater use for crops (GWC) (in MAF/BCM) as well as the major direct

15 energy uses in agriculture (in 1000 total oil equivalents or kTOE): energy use for groundwater

16 pumping (EnGW) and energy use for crop production (EnC, which is energy use for tractors) under

17 different temperature and precipitation combinations in a “climate response table.” Baseline values

18 of these four resource use are also listed in the table. Compared to Table S2, temperature increases

19 of 1.5, 3 and 4.5 oC can be approximately treated as the near, middle and far future.

20 Results indicate that annual mean SWC is sensitive to both temperature and precipitation

21 changes but the magnitude of changes is smaller compared with GWC and EnGW. This is possibly

22 due to the highly constrained canal water allocation system that limits changes in surface water

23 use but does not affect groundwater use. GWC is very sensitive to temperature changes and is

10
 1 highly correlated (almost linearly) with EnGW. Among these four parameters, EnGW experiences

2 the largest changes while EnC changes the least. In general, a hotter climate results in increased

3 water and energy use while a wetter climate results in larger surface water use (due to expansion

4 of crop area in the model), which slightly tampers increases of groundwater and energy use. Large

5 fluctuations of resource use as a result of inter-annual variability are important concerns for the

6 management of these resources. Contrary to the annual mean pattern, hotter and wetter climate

7 conditions will decrease the inter-annual variability of SWC. Unlike the annual mean, under drier

8 climate, GWC is sensitive to both precipitation and temperature changes. EnGW remains sensitive

9 to temperature change and shows no clear relationship to GWC’s inter-annual variability. EnC is

10 more sensitive to precipitation than temperature changes. In general, a hotter climate will result in

11 large variabilities of groundwater and energy uses but will reduce the variability of surface water

12 uses. A wetter future, on the other hand, will reduce variability of SWC and GWC and can mitigate

13 the increasing trend of EnGW and EnC caused by increasing temperatures.

14 4.1.2. Impact of climate change, dam investments and water allocation policies on resource use

15 This section describes the combined effects of all scenarios: climate change, new water

16 infrastructure and alternative water allocation policies. Parallel coordinate plots are used to present

17 the multiple dimensional outcomes where the first four axes denote alternative scenarios and the

18 latter axes reflect resource use and basin outputs. We apply the ex-post scenario identification

19 process (Groves and Lempert, 2007) and define long-term averages of resource use, crop

20 production and hydropower generation under historical conditions as thresholds. These thresholds

21 are used to define “acceptable” (thin green lines) and “unacceptable” (bold brown lines) scenarios

22 for the basin. Thus, the management purpose here is to maintain historical basin outputs

23 (thresholds become the lower bound) while using less water and energy resources (thresholds

24 become the upper bound). These thresholds are chosen for illustrative purposes. Alternative

11
 1 thresholds, such as targets for increased energy efficiency or agricultural production or reduced

2 water use could also be specified based on input from decision makers in Pakistan.

3 Figure 2 shows the result of annual basin-wide resource use under all scenarios. The fifth

4 to eight axes denote values for SWC, GWC, EnGW and EnC, respectively. Results show that

5 changes in water allocation policy is the dominant driver affecting resource use. When the system

6 is using the “Historical” water allocation, only 2% of the basin scenarios are “unacceptable.”

7 “Basin” and “IRSA” water allocation rules result in 31% and 78% of scenarios having values

8 higher the historical results. In general, SWC increases when water allocation changes from

9 “Historical” to “IRSA” or “Basin.” On the contrary, GWC and EnGW decrease when water

10 allocation policies change. A more flexible surface water allocation policy will reduce energy use

11 but increases surface water use in the basin. Therefore, the tradeoff relationship between surface

12 water and groundwater can be seen as a tradeoff between available surface water and available

13 energy for groundwater pumping. Meanwhile, EnC will also increase when water allocation

14 changes from “Historical” to “IRSA” to “Basin” because more crops are being grown. Decreasing

15 precipitation and increasing temperature will result in fewer “acceptable” scenarios, as expected.

16 At the most extreme case, declines in precipitation by 20% and temperature increases of 4.5 oC

17 will result in 41% and 64% of scenarios that are deemed “unacceptable,” respectively. Combining

18 results from Table 1 and Figure 2, we also note that while climate change increases EnGW and

19 EnC, a more flexible water allocation policy can mitigate increasing energy uses.

20 4.1.3. Crop production and hydropower generation

21 The effects of different scenarios on basin-wide production of sugarcane (a representative

22 cash crop) and wheat (a representative food crop) as well as hydro-electricity generation are shown

23 in Figure 3. Scenario results are deemed “unacceptable” when all evaluated basin outputs are

24 below the historical long-term average. Results show that decreasing precipitation and increasing

12
 1 temperature are the two dominant drivers for “unacceptable” basin scenarios. New water

2 infrastructures and flexible water allocation policies can mitigate these negative climate change

3 impacts. For example, the “Historical” system accounts for 17% of “unacceptable” basin scenarios

4 while the Mangla rising project (“New Mangla”) reduces that number to 9%. Adding Diamer-

5 Basha Dam (“New Mangla plus Basha”) will ensure that all scenarios are “acceptable,” that means

6 historic output levels can be met even under higher temperatures and decreased precipitation levels.

7 Similarly, “IRSA” and “Basin” water allocation can satisfy thresholds that the “Historical” basin

8 water allocation would violate. Note that these results would change if we select different basin-

9 level outputs such as different crops or we disaggregate results at the province level. The latter

10 will be discussed in the following section.

11 Table 2 presents impacts of climate change, dam investment and changes in water

12 allocation policies on water and energy use per unit of output. This table summarizes changes on

13 the nexus between water, energy and food. The first half of Table 2 presents climate change

14 impacts on WEFN in the historical condition. For example, under a precipitation decrease of 20%

15 combined with a temperature increase of 4.5oC, 35% more energy is needed for groundwater

16 pumping (7.35 to 9.96 kTOE/MAF or 5.96 to 8.08 TOE/BCM), 36% more energy is needed for

17 crop production (7.49 to 10.18 kTOE/Mton) and 15% more water is needed for crop production

18 (1.28 to 1.47 MAF/Mton or 1.57 to 1.81 BCM/Mton). The lower half of Table 2 presents an

19 example of how new water infrastructure together with alternative water allocation policies can

20 help mitigate growing resource use. For illustrative purposes, the combination of “New Mangla

21 plus Basha” and the “Basin” water allocation policy is shown. Results indicate that energy use per

22 unit of crop produced and groundwater pumping can fall below baseline levels under some climate

23 change-infrastructure-water policy scenarios while water use per unit of crop produced would

13
 1 remain above baseline levels. For example, under a precipitation decrease of 20% combined with

2 a temperature increase of 4.5oC, incremental energy needs for groundwater pumping decrease from

3 35% (9.96 kTOE/MAF or 8.08 kTOE/BCM) under the historical scenario to 14% (8.42

4 kTOE/MAF or 6.83 kTOE/BCM) and additional energy needs for crop production decrease from

5 36% (10.18 kTOE/Mton) to 10% (8.26 kTOE/Mton) over a non-climate change baseline scenario.

6 But water use either stays the same or slightly increases. This confirms the result from the previous

7 section that flexible surface water allocation policy can reduce energy use but would not conserve

8 on water use.

9 4.2. Provincial level results

10 4.2.1. Impact of climate change on resource use

11 Provincial-level results are as important as basin-wide output because key water, energy

12 and food production decisions are taken by provincial-level decision makers. We focus here on the

13 two provinces that are critical for agricultural production: Punjab and Sindh. As we saw divergent

14 developments between surface water use and energy use for groundwater pumping, this section

15 focuses on these two resource use parameters.

16 Table 3 presents impacts of climate change on annual mean and inter-annual variability of

17 resource use for two provinces. Annual mean SWC has a similar pattern in both provinces: hotter

18 and wetter climate increases SWC and drier and warmer climate decreases SWC. However, in

19 Punjab surface water use responds particularly strongly under drier and warmer climate while the

20 response in Sindh is larger under hotter and wetter climate. Annual mean energy use for

21 groundwater in Punjab is sensitive to temperature while sensitivity in Sindh relates to precipitation.

22 This is likely because Punjab as an upstream province pumps groundwater primarily based on crop

23 water requirements (which is largely controlled by temperature) while Sindh as a downstream

14
 1 province pumps groundwater to compensate for a lack of surface water availability. The changes

2 are again larger for EnGW than for SWC in both provinces. In terms of inter-annual variability,

3 changes in SWC and EnGW in Punjab are a reflection of basin-wide results. Inter-annual

4 variability of SWC is sensitive to both temperature and precipitation and has an opposite pattern

5 as the annual mean. Inter-annual variability of EnGW is sensitive to temperature like the annual

6 mean of the same value. Inter-annual variability of SWC and EnGW in Sindh, on the other hand,

7 follows the same pattern. Their relative use both decreases under hotter and wetter climate

8 conditions.

9 4.2.2. Impact of climate change, dam investments and water allocation policies on resource use

10 Figure 4 presents impacts of combined scenarios on provincial-level resource use focusing

11 on the diverging parameters of SWC and EnGW only. Defining “unacceptable” provincial-level

12 scenarios is more complicated and depends on the interpolation of model results for provincial-

13 level policy. Following the basin-wide setting, we define water use above long-term averages as

14 “unacceptable” scenarios. However, SWC describes actual surface water use, it can be seen as the

15 actual amount of water each province obtains. Therefore, one can compare that value with the

16 assigned water right and make the argument (in our definition) that a value below the water right

17 is “unacceptable.” All in all, when the modeling results are used for different policy dialogues,

18 these thresholds and “unacceptable” scenarios should be specified based on the common interests

19 of different policy makers.

20 Based on our definition, brown lines in Figure 4 describe scenarios where both Punjab and

21 Sindh use more water and energy compared to long-term average use. From a management

22 perspective, such brown lines could form a basis for negotiations between the two provinces.

23 Results indicate that “Historical” water allocation would result in 20% “unacceptable” provincial

24 scenarios while “IRSA” and “Basin” water allocation would not result in any “unacceptable”

15
 1 provincial-level scenarios. Both “Historical” and “New Mangla” would have about 10%

2 “unacceptable” provincial scenarios and investment in Diamer-Basha Dam would reduce this

3 number to 2%. Increasing temperature will result in more “unacceptable” provincial scenarios,

4 similar to the basin-wide result. For example, a 4.5 oC temperature increase will render 22% of

5 provincial scenarios “unacceptable.” Although not shown in the figure, model results also suggest

6 that changes in surface water allocation will affect the tradeoff between SWC and EnGW at the

7 provincial level. Under the “IRSA” allocation, when Punjab uses less SWC, Sindh will use more.

8 Correspondingly, Sindh would use less EnGW and Punjab would use more. An interactive version

9 of web-based parallel coordinate plots has been developed allowing users to specify the impact of

10 climate change and human activities on both basin-wide and provincial level resource use

11 (http://people.umass.edu/yceyang/IndusWEF.html). Readers are encouraged to explore this

12 website and use the direct brush function to highlight the range of climate change and human

13 activities to see their impacts on water/energy uses.

14 4.2.3. Cash and food crop production

15 Figure 5 shows the annual production of cash (sugarcane) and food (wheat) crops in these

16 two provinces. “Unacceptable” provincial-level scenarios are defined as those where crop

17 production is below the long-term average. Both cash and food crops are affected by climate

18 change under the “Historical” water allocation. Additional storage through Mangla and Diamer-

19 Basha dams do not significantly reduce adverse climate change impacts on food production.

20 “IRSA” and “Basin” water allocation, on the other hand, successfully address “unacceptable”

21 provincial-level scenarios, but have their own drawbacks. “IRSA” water allocation would reduce

22 sugarcane production in Sindh and the “Basin” allocation policy would inevitably encounter

23 political and social obstacles if implemented as the province-level accord would need to be

24 adjusted to accommodate this allocation mechanism.

16
 1 5. Discussion

2 Results presented in the previous section focus on long-term water and energy uses and

3 related food and hydropower generation. In the following two sections, we discuss inter-year

4 trends which were ignored in the result section for selected resources and the modeling results

5 from this Indus case study can be used to inform more general WEFN issues in other places in the

6 world, respectively.

7 5.1. Inter-year trend analysis of water and energy uses

8 All previously presented results are for long-term averages with no information on year-

9 by-year patterns of water and energy uses. In this section, we use the Mann-Kendall trend test to

10 identify significant trends, if any, in resource use patterns. The Mann-Kendall trend test is one of

11 the most widely used non-parametric tests to detect significant trends in time series (Hamed, 2008).

12 Table 4 shows the results of this trend test under no climate change impacts. Negative coefficients

13 describe a downward trend and positive coefficients describe upward trends. The star markers

14 represent statistical significance at the 95% confidence interval.

15 EnGW has a strong upward trend in Punjab under all infrastructure development scenarios.

16 This means more energy is used for groundwater pumping as a result of declining water tables.

17 Since Punjab’s groundwater pumping represents 95% of basin-wide pumping for agriculture,

18 basin-wide EnGW shows the same strong upward trend. Energy use for tractors shows a declining

19 trend (caused by a slight decline of crop production due to expensive groundwater pumping over

20 time). Unlike trends in energy use, trends in water use are largely influenced by surface water

21 allocation policies. If the system follows the “Historical” or “IRSA” allocation, no significant trend

22 is found for SWC but Punjab’s (and therefore basin-wide) GWC shows a downward trend. This is

23 because pumping groundwater becomes costly over time. If the system follows the “Basin”

17
 1 allocation policy, Sindh and basin-wide SWC follow a downward trend while GWC in Sindh

2 follows an upward trend. This suggests that the downstream province needs to pump more

3 groundwater to compensate for decreasing surface water availability (due to upstream province

4 relies on more surface water cause by expensive groundwater pumping) over time.

5 5.2. Knowledge transfer from the Indus River case study to other places

6 WEFN results highlighted in this Indus River study have the potential to inform other

7 similar environments in the world with arid to semi-arid climate and surface water supply from

8 snow/glacier melt like California and Central Asia. For example, hotter climate will result in more

9 water and energy use which might not be an immediate threat due to increasing water supply from

10 glacier melting but it will be a significant long-term concern after glaciers are completely melted

11 in these systems. When new policies and/or infrastructure are introduced into the system (e.g. the

12 new “Sustainable Groundwater Management Act” in California and the ongoing Rogun Dam

13 debate in Central Asia), flexible water allocation policy can reduce tradeoffs between surface water

14 and energy uses and also mitigate adverse climate change impacts. Finally, climate change will

15 likely increase nexus in all three aspects: energy use per unit of water, energy use per unit of crop

16 produced and water use per unit of crop produced. New water infrastructure might mitigate climate

17 change impacts on energy use but is unlikely to reduce water use. This implies that a portfolio of

18 solutions with nexus thinking (e.g. new water infrastructure combined with water-saving irrigation

19 or water-saving crop technologies) is needed to simultaneously address water, energy, and food

20 security issues in these systems.

21 There are several future research opportunities to further expand the current modeling

22 structure and apply it to other systems. First, the groundwater module is relatively simple

23 compared to the surface water module in the current model. Linking with a physically-based

18
 1 groundwater model can reduce underground water balance uncertainties, improve the estimation

2 of both GWC and EnGW and better address groundwater regulations, such as the new legislation

3 in California. The definition of WEFN should be expanded to beyond the agricultural sector. A

4 linkage with an integrated energy model with regional specification that considers domestic,

5 commercial and industrial energy markets would provide better understanding of the overall

6 WEFN in the system. Finally, demographic dynamics affect food and energy demands, which is a

7 concern in most developing countries and should be explicitly addressed in the expanded modeling

8 structure.

9 6. Conclusion

10 Satisfying rapidly growing food, water and energy demands will be a challenging task

11 globally and particularly in Pakistan. Insufficient food production has been reported since 2003

12 and the country is also suffering from growing energy deficits since 2005. Pakistan is eager to

13 develop new energy sources do address the energy crisis while maintaining or increasing food

14 supplies. Hydropower generation is considered one of the cheapest renewable energy sources for

15 Pakistan but water uses for hydropower generation in the Indus Basin might well compete with

16 water uses for crop production, which is essential for basic food security, employment and water

17 supply. Therefore, an understanding of the connections of water, energy and food in the basin with

18 nexus thinking is critical for the country’s food and energy development plans. This study applied

19 a systems analysis modeling approach to evaluate impacts of climate change as well as alternative

20 dam investments and water allocation policies on energy and water use for crop production and

21 hydropower generation.

22 We find that annual water and energy uses at both the basin and provincial levels will

23 increase with hotter and wetter climate. Groundwater use for crops (GWC) and energy use for

19
 1 pumping (EnGW) are in general more sensitive to temperature increases than precipitation

2 increases. The inter-annual variability of water uses will decrease when both temperature and

3 precipitation increase. On the other hand, the inter-annual variability of energy use has an opposite

4 pattern and will increase with temperature and precipitation increase.

5 Energy use per unit of crop produced is expected to increase by 36% under a more extreme

6 climate change scenario (combining a temperature increase of 4.50C with a decrease in

7 precipitation by 20%). Almost similar growth (35%) is projected for energy requirements per unit

8 of groundwater and a smaller growth (13%) for water use per unit of crop produced is expected

9 under the historical water allocation system. Flexible surface water allocation policies can more

10 effectively mitigate negative climate change impacts on resource use than new water infrastructure.

11 However, flexible surface water allocation policies will also result in increased water use. At the

12 provincial level, the tradeoff between surface water for crop (SWC) and EnGW is clearer and the

13 pattern changes with different water allocation policies. Under the “IRSA” allocation, Punjab

14 would use more SWC, less EnGW and Sindh would use more EnGW and less SWC. An opposite

15 pattern has been observed when the “Basin” allocation method is adapted. Under this practice,

16 Punjab would use less SWC, more EnGW and Sindh would use less EnGW and more SWC.

17 The inter-year trend analysis of water and energy uses shows that energy uses have a clearer

18 trend over time than water uses. EnGW has a strong upward trend due to the water table drawdown

19 while energy uses for tractors (EnC) show a downward trend. Surface water allocation will affect

20 the inter-year water use trends in Punjab and Sindh. The “Basin” allocation will result in an upward

21 trend in GWC in Sindh due to the decreasing surface water availability. Results highlighted in this

22 Indus Basin study can be used as a reference to inform the WEFN in other systems like California

23 and Central Asia with similar hydro-climatic environments. Future studies should link with a

20
 1 detailed groundwater model and an energy market model while explicitly address the food and

2 energy demand changes caused by population growth.

3 Acknowledgement

4 The paper was supported by USAID through the International Food Policy Research

5 Institute (IFPRI)-led Pakistan Strategy Support Program under the CGIAR Research Program on

6 Policies, Institutions and Markets. The views expressed in this paper are those of the authors and

7 do not necessarily reflect the views of USAID or IFPRI. Authors would like to thank the editor,

8 associate editor and two anonymous reviewers for their positive comments on an earlier version

9 of this manuscript.

10 References

11 Ahmad, I. 2009. Optimal Control of Multiple reservoirs system under water scarcity, Ph.D.

12 dissertation, University of the Punjab, Lahore, Pakistan.

13 Amir P. 2005. The Role of Larger Dam in the Indus Basin, Country Water Resources Assistance

14 Strategy Background paper #10, The World Bank, Washington D.C., USA.

15 Bhutta, Z. (2011) Pakistan National Nutrition Survey, Aga Khan University, Pakistan.

16 Biggs E. M. et al. 2015. Sustainable development and the water–energy–food nexus A perspective

17 on livelihoods. Environmental Science and Policy, 54: 389-397.

18 Briscoe, J. and Qamar, U. eds. 2006. Pakistan’s Water Economy: Running Dry. Washington:

19 World Bank, Washington, D.C.

20 Brown, C., Y. Ghile, M. Laverty, and K. Li. 2012. Decision scaling: Linking bottom-up

21 vulnerability analysis with climate projections in the water sector, Water Resour. Res., 48,

22 W09537, doi:10.1029/2011WR011212.

21
 1 Duloy, J. H. and O’Mara, G. T. 1984. Issues of Efficiency and Interdependence in water resource

2 investments: lessons from the Indus basin of Pakistan. Washington: World Bank,

3 Washington, D.C.

4 Energy Information Administration (EIA). 2008. http://www.eia.gov/countries/data.cfm on

5 2016/04/11.

6 Groves, D. G. and R. J. Lempert 2007. A new analytic method for finding policy-relevant scenarios.

7 Global Environmental Change, 17:73-85.

8 Hamed, K. H. 2008. Trend detection in hydrologic data: The Mann Kendall trend test under the

9 scaling hypothesis. Journal of Hydrology, 349 (3-4): 350-363.

10 Hussain, S. A. 1993, Hydropower for Pakistan, IEE Review, 39(3):123-125.

11 International Research Group (IRG). 2011. Pakistan Integrated Energy Model (Pak-IEM), Final

12 Report, ADB TA No. 4982-PAK, Islamabad, Pakistan.

13 Irshad, R. 2011. Water Apportionment Accord 1991, Indus River System Authority.

14 http://www.uaf.edu.pk/golden_jubilee/downloads/presentations/Rao Irshad 10.01.11-

15 Chairman IRSA.ppt accessed at 2016/04/11.

16 Izhar-ul-Haq and Munir, K. 2009. Role of Hydropower in Management of Power Crisis in Pakistan,

17 Symposium on “Genesis of Power Crisis and its Management in Pakistan,” 31: 19-34.

18 Khan, N. M. and Tingsanchali, T. 2009. Optimization and Simulation of Reservoir Operation with

19 Sediment Evacuation: A Case Study of the Tarbela Dam, Pakistan, Hydrology Process,

20 23:730-747.

21 Luna, B. A. and Jabbar, M. 2011. Kalabagh - A Superior dam designer’s view point, Pakistan

22 Engineering Congress, 71st Annual Session Proceedings, 287-302.

22
 1 Mekonnen, D. A. Siddiqui and C. Ringler. 2016. Drivers of groundwater use and technical

2 efficiency of groundwater, canal water, and conjunctive use in Pakistan’s Indus Basin

3 Irrigation system. International Journal of Water Resources Development.

4 10.1080/07900627.2015.1133402.

5 Mirza, U. K., Ahmad, N., Majeed, T. and Harijan, K. 2008. Hydropower Use in Pakistan: Past

6 Present and Future, Renewable and Sustainable Energy Reviews, 12:1641-1651.

7 NESPAK. 2013. Upgrading of tools, water resources, database management systems and models

8 - Final report: IBMR, GIS and database integration (WCAP-B1), Lahore, Pakistan.

9 Planning Commission, 2014. Pakistan2025: One nation, one vision. Planning Commission,

10 Islamabad,Pakistan. http://www.pc.gov.pk/wp-content/uploads/2015/05/Pakistan-Vision-

11 2025.pdf accessed on 2016/04/11.

12 Qureshi, A. S, Shah, T. and Akhtar, M. 2003. The groundwater economy of Pakistan. Working

13 Paper 64. International Water Management Institute. Lahore, Pakistan.

14 http://www.pildat.org/publications/publication/WaterR/Inter-

15 ProvincialWaterIssuesinPakistan-BackgroundPaper.pdf accessed on 2016/04/11.

16 Räsänen, T. A. Joffre, O. M., Someth, P., Thanh, C. T. Keskinen, M. and Kummu, M. 2014.

17 Model-Based Assessment of Water, Food, and Energy Trade-Offs in a Cascade of

18 Multipurpose Reservoirs: Case Study of the Sesan Tributary of the Mekong River. Journal

19 of Water Resources Planning and Management - ASCE, 10.1061/(ASCE)WR.1943-

20 5452.0000459, 05014007.

21 Ringler and Anwar, 2013. Water for food security—Challenges for Pakistan. Water International

22 38(5): 505-514.

23
 1 Scott C. A., Pierce, S. A., Pasqualetti, M. J., Jones, A. L., Montz, B. E. and Hoover, J. H. 2011.

2 Policy and institutional dimensions of the water–energy nexus. Energy Policy, 39: 6622-

3 6630.

4 Siddiqi, A., Wescoat, J., L. Jr, Humair, S. and Afridi, K. 2012. An Empirical Analysis of the

5 Hydropower Portfolio in Pakistan, Energy Policy, 50: 228-241.

6 Tate E. L. and Farquharson, F. A. K. 2000. Simulating Reservoir Management under the Threat

7 of Sedimentation: The Case of Tarbela Dam on the River Indus, Water Resources

8 Management 14: 191–208.

9 WAPAD, 2011a. Hydro potential in Pakistan, Lahore, Pakistan.

10 WAPAD, 2011b. Tarbela 4th Extension Hydropower Project Environmental and Social

11 Assessment, Lahore, Pakistan.

12 WAPDA 2012 Mangla dam raising project http://www.wapda.gov.pk/pdf/manglar1010.pdf.

13 accessed on 2016/04/11.

14 Winston, A. 2013. The Water-Food-Energy Nexus. The Initiative for Global Environmental

15 Leadership Conference-Workshop: The Nexus of Energy, Food and Water. Philadelphia,

16 PA.

17 Yang, Y. C. E., Brown, C. M., Yu, W. H. and Savitsky, A. 2013. An Introduction to IBMR – A

18 Hydro-Economic Model for the Climate Change Impact Assessment in the Indus River in

19 Pakistan. Water International, 38(5): 632-650.

20 Yang, Y. C. E., Brown, C. M., Yu, W. H., Wescoat, J. L. Jr. and Ringler, C. 2014. Water

21 Governance and Adaptation to Climate Change in the Indus River Basin. Journal of

22 Hydrology, 519: 2527-2537.

24
1 Yang Y. C. E., Wi, S., Ray, P. A., Brown C. M. and Khalil, A. 2016. The Future Nexus of the

2 Brahmaputra Basin: Climate, Water, Energy and Food trajectories, Water Resources

3 Research, 37: 16-30.

4 Yu, W., Yang, Y. C. E., Savitsky, A., Alford, D., Brown, C. Wescoat, J., Debowicz, D. and

5 Robinson, S. 2013. The Indus Basin of Pakistan: The Impacts of Climate Risks on Water

6 and Agriculture. Washington, DC: World Bank. doi: 10.1596/978-0-8213-9874-6. License:

7 Creative Commons Attribution CC BY 3.0

25
Table Click here to download Table Indus-water-energy-Table-
0425.docx

Table 1. Climate change impacts on basin-wide water (in MAF = 1.233 billion cubic meters) and
energy (kTOE) use changes (historical system and historical water allocation)
Annual mean Inter-annual variability*
Surface water Precipitation (%) Surface water Precipitation (%)
uses for crops -20 -10 0 10 20
uses for crop -20 -10 0 10 20
(107 MAF or 131 BCM) (3.2 MAF or 3.94 BCM)
4.5 0.58 2.02 2.80 3.72 3.75 4.5 -10.88 -7.29 -26.68 -46.89 -63.68
Temperature Temperature
3.0 -3.03 -0.46 1.01 1.60 3.55 3.0 32.64 24.63 -33.68 -35.14 -30.89
(oC) (oC)
1.5 -4.45 -1.61 -1.00 -0.30 1.31 1.5 60.29 5.78 -9.74 -14.27 -29.41
Groundwater use Precipitation (%) Groundwater use Precipitation (%)
for crops for crop
(51 MAF or 62.9 BCM)
-20 -10 0 10 20 (0.6 MAF or 0.74 BCM)
-20 -10 0 10 20
4.5 33.50 33.03 33.18 32.28 31.27 4.5 86.63 31.40 39.65 -1.94 -14.43
Temperature Temperature
3.0 18.60 16.97 16.29 15.73 15.36 3.0 186.51 75.66 37.12 26.25 7.07
(oC) (oC)
1.5 9.36 8.01 5.33 6.03 6.01 1.5 206.35 88.11 16.88 -24.07 -11.03
Direct energy Precipitation (%) Direct energy Precipitation (%)
uses for uses for
groundwater -20 -10 0 10 20 groundwater -20 -10 0 10 20
(375 kTOE) (112 kTOE)
4.5 80.85 76.46 73.43 70.75 68.47 4.5 122.74 115.34 108.65 105.24 101.58
Temperature Temperature
3.0 48.30 41.22 37.05 34.39 32.00 3.0 74.69 61.57 54.12 49.44 45.46
(oC) (oC)
1.5 29.09 21.68 23.28 14.26 12.12 1.5 47.41 33.66 23.92 21.63 16.81
Direct Energy Precipitation (%) Direct energy Precipitation (%)
use for crops (550 uses for crop (2.3
kTOE)
-20 -10 0 10 20 kTOE)
-20 -10 0 10 20
4.5 -1.97 -1.91 -1.86 -1.85 -1.85 4.5 95.90 89.38 98.09 79.06 74.80
Temperature Temperature
3.0 -1.21 -1.15 -0.97 -0.93 -0.90 3.0 98.64 70.34 74.71 68.95 66.26
(oC) (oC)
1.5 -0.62 -0.55 -0.85 -0.08 -0.39 1.5 93.49 36.60 50.03 21.09 29.52
*
Standard deviation over 30-year simulation period is used to represent inter-annual variability

1
Table 2. Changes of annual mean for “energy use per unit of groundwater” (kTOE per MAF or BCM) and “water/energy use per unit
of crop” (MAF or BCM per Mton) under climate change impacts
Precipitation change (%)
-20 -10 0 10 20
energy use per unit of Temperature 4.5 9.96/8.08* 9.75/7.91 9.58/7.77 9.49/7.70 9.44/7.66
(ground)water; (baseline = increase 3.0 9.19/7.45 8.88/7.20 8.67/7.03 8.54/6.93 8.41/6.82
7.35 kTOE per MAF or 5.96
kTOE per BCM)
(oC) 1.5 8.68/7.04 8.28/6.72 8.61/6.98 7.92/6.42 7.78/6.31
Historical system Temperature 4.5 10.18 10.04 9.93 9.85 9.78
under “Historical” energy use per unit of crop increase 3.0 9.07 8.84 8.71 8.62 8.55
(baseline = 7.49 kTOE/Mton) o
water allocation ( C) 1.5 8.45 8.20 8.25 7.96 7.89
water use per unit of crop Temperature 4.5 1.47/1.81+ 1.48/1.82 1.48/1.82 1.49/1.84 1.49/1.84
(baseline = 1.28 MAF/Mton or increase 3.0 1.36/1.68 1.37/1.69 1.38/1.70 1.38/1.70 1.39/1.71
1.57 BCM/Mton) (oC) 1.5 1.30/1.60 1.31/1.62 1.31/1.62 1.31/1.62 1.32/1.63
energy need per unit of Temperature 4.5 8.42/6.83* 8.15/6.61 8.03/6.51 7.75/6.29 7.72/6.26
(ground)water increase 3.0 7.72/6.26 7.38/5.99 7.21/5.85 7.00/5.68 6.83/5.54
(oC) 1.5 7.04/5.71 6.94/5.63 6.65/5.39 6.47/5.25 6.24/5.06
New Mangla plus Temperature 4.5 8.26 8.07 7.94 7.79 7.69
energy use per unit of crop
Basha under “Basin” increase 3.0 7.48 7.23 7.10 6.98 6.90
water allocation (oC) 1.5 6.94 6.74 6.60 6.50 6.43
Temperature 4.5 1.48/1.82+ 1.50/1.85 1.50/1.85 1.51/1.86 1.51/1.86
water use per unit of crop 3.0
increase 1.34/1.65 1.36/1.68 1.37/1.69 1.38/1.70 1.40/1.73
(oC) 1.5 1.28/1.58 1.30/1.60 1.30/1.60 1.31/1.62 1.31/1.62
*
per MAF/per BCM; +MAF per Mton/BCM per Mton

2
 Table 3. Changes (water in MAF and energy in kTOE) in provincial surface water use for crop
and energy use for groundwater, baseline (historical system and historical water allocation) and
under climate change impacts
Annual mean Inter-annual variability*
Surface water uses Precipitation (%) Surface water Precipitation (%)
for crop -20 -10 0 10 20
uses for crop -20 -10 0 10 20
(61.2 MAF or 75.4 BCM) (1.8 MAF or 2.2 BCM)
4.5 -1.75 0.16 0.55 1.71 1.51 4.5 17.47 8.72 -39.14 -2.22 -57.51
Temperature Temperature
3.0 -4.94 -1.42 -0.65 0.21 1.40 3.0 94.69 47.53 -17.87 -37.81 -30.12
(oC) (oC)
Punjab

1.5 -6.43 -3.28 -0.90 -0.47 -0.06 1.5 132.59 36.34 43.93 -32.88 -49.57
Energy use for Precipitation (%) Energy use for Precipitation (%)
groundwater groundwater
(369 kTOE)
-20 -10 0 10 20 (114 kTOE)
-20 -10 0 10 20
4.5 81.20 76.80 73.80 71.11 68.90 4.5 120.97 113.73 107.15 103.76 100.05
Temperature Temperature
3.0 48.45 41.49 37.24 34.62 32.29 3.0 73.56 60.67 53.39 48.67 44.85
(oC) (oC)
1.5 29.04 21.72 23.39 14.32 12.17 1.5 47.00 33.37 23.74 21.35 16.60
Surface water uses Precipitation (%) Surface water Precipitation (%)
for crop -20 -10 0 10 20
uses for crop -20 -10 0 10 20
(38.6 MAF or 47.6 BCM) (2.2 MAF or 2.71 BCM)
4.5 3.34 4.16 5.61 6.20 6.58 4.5 -28.39 -9.40 -15.67 -48.52 -61.60
Temperature Temperature
3.0 -0.67 0.50 3.41 3.49 6.81 3.0 -4.55 12.49 -30.37 -11.64 -16.83
(oC) (oC)
Sindh

1.5 -1.95 0.51 -1.53 -0.50 3.13 1.5 35.42 -6.51 -13.88 17.82 0.08
Energy use for Precipitation (%) Energy use for Precipitation (%)
groundwater groundwater
(3.8 kTOE)
-20 -10 0 10 20 (1.1 kTOE)
-20 -10 0 10 20
4.5 26.69 22.37 15.87 11.72 1.88 4.5 -13.74 -13.60 -19.56 -18.69 -29.55
Temperature Temperature
3.0 23.32 7.01 8.33 1.01 -8.59 3.0 -13.11 -17.34 -9.07 -20.64 -10.30
(oC) (oC)
1.5 26.69 14.62 11.86 3.46 0.48 1.5 13.58 11.22 11.08 -0.89 -1.05
*
Standard deviation over 30-year simulation period is used to represent inter-annual variability

3
Table 4. Mann-Kendall trend analysis for water and energy uses under different infrastructure
developments
System Allocation Regional SWC GWC EnGW EnC
Basin -0.19 -0.39* 1.00* -0.94*
Historical Punjab 0.08 -0.60* 1.00* -0.94*
Sindh -0.44* 0.65* 0.07 0.70*
Basin -0.33* -0.63* 0.99* -0.91*
Historical IRSA Punjab 0.20 -0.81* 0.99* -0.91*
Sindh 0.20 0.18 -0.53* 0.33*
Basin -0.34* 0.21 0.90* -0.79*
Basin Punjab -0.13 0.02 0.91* -0.81*
Sindh -0.55* 0.40* -0.17 -0.04
Basin -0.28* -0.32* 1.00* -0.99*
Historical Punjab 0.05 -0.71* 1.00* -0.99*
Sindh -0.39* 0.68* 0.21 0.64*
Basin -0.26 -0.69* 0.99* -0.92*
New Mangla IRSA Punjab 0.20 -0.83* 0.99* -0.93*
Sindh 0.20 0.58* -0.34* 0.15
Basin -0.32* 0.25 0.89* -0.77*
Basin Punjab -0.08 0.01 0.89* -0.78*
Sindh -0.49* 0.34* 0.00 0.05
Basin -0.09 -0.56* 1.00* -0.95*
Historical Punjab 0.05 -0.71 1.00* -0.95*
Sindh -0.13 0.23 -0.63* 0.06
Basin -0.18 -0.58* 0.99* -0.92*
New Mangla plus Basha IRSA Punjab 0.20 -0.81* 0.99* -0.94*
Sindh 0.20 0.60* -0.01 0.27
Basin -0.44* 0.28* 0.89* -0.79*
Basin Punjab -0.18 0.02 0.89* -0.79*
Sindh -0.47* 0.57* -0.44* 0.71*
*
means significant under 95% confidence interval

4
Figure Click here to download Figure Figure 1.pdf

Figure 1. The location of current and potential reservoirs and agro-climatic zones by
province in the Indus River of Pakistan

1
Figure Click here to download Figure Figure 2.pdf

Figure 2. Basin-wide annual water/energy uses under the impacts of climate change (first two axes), dam development (third axis) and
alternative water allocation scenarios (fourth axis). Green lines represent scenarios that satisfy thresholds and brown lines
represent scenarios that cannot satisfy the thresholds. In this figure, scenarios that result in all four water/energy uses above
historical conditions are labeled as brown lines.

2
Figure Click here to download Figure Figure 3.pdf

Figure 3. Basin-wide annual crop and hydropower production under the impacts of climate change (first two axes), dam development
(third axis) and alternative water allocation scenarios (fourth axis). Green lines represent scenarios that satisfy thresholds and
brown lines represent scenarios that cannot satisfy threshold. In this figure, scenarios that result in sugarcane production, wheat
production and hydropower generation all below historical condition are labeled as brown lines.

3
Figure Click here to download Figure Figure 4.pdf

Figure 4. Punjab and Sindh annual water/energy uses under the impacts of climate change (first two axes), dam development (third
axis) and alternative water allocation scenarios (fourth axis). Green lines represent scenarios that satisfy thresholds and brown
lines represent scenarios that cannot satisfy threshold. In this figure, scenarios that result in all water/energy uses in both
Punjab and Sindh above historical condition are labeled as brown lines.

4
Figure Click here to download Figure Figure 5.pdf

Figure 5. Punjab and Sindh crop production under the impacts of climate change (first two axes), dam development (third axis) and
alternative water allocation scenarios (fourth axis). Green lines represent scenarios that satisfy thresholds and brown lines
represent scenarios that cannot satisfy threshold. In this figure, scenarios that result in sugarcane and wheat production in both
provinces below historical condition are labeled as brown lines.

View publication stats