Analysis of Operational Changes of Tarbela Reservoir to Improve the Water Supply, Hydropower Generation, and Flood Control Objectives

sustainability

Article Analysis of Operational Changes of Tarbela Reservoir to Improve the Water Supply, Hydropower Generation, and Flood Control Objectives

Ahmed Raﬁque 1,*, Steven Burian 1, Daniyal Hassan 1 and Rakhshinda Bano 2

1 Department of Civil and Environmental Engineering, University of Utah, Salt Lake City, UT 84112, USA; [email protected] (S.B.); [email protected] (D.H.) 2 U.S.-Pakistan Center for Advanced Studies in Water, Mehran University of Engineering and Technology, Jamshoro City, Sindh 76062, Pakistan; [email protected] * Correspondence: ahmed.raﬁ[email protected]

Received: 30 August 2020; Accepted: 14 September 2020; Published: 22 September 2020

Abstract: In this study, a model was created with the Water Evaluation and Planning (WEAP) System and used to explore the benefits of altering the operations of Tarbela Dam in terms of reliability, resilience, and vulnerability (RRV) for the three objectives of irrigation supply, hydropower generation, and flood control. Sensitivity analysis and logical reasoning with operators identified a feasible operational rule curve for testing using the integrated performance analysis. The reservoir performance for the altered operations was compared to the baseline performance following current operations for both historical and projected future climate and water demand conditions. Key simulation results show that the altered operations strategy tested under historical climate and water demand conditions would increase RRV by 17%, 67%, and 7%, respectively, for the water supply objective and 34%, 346%, and 22%, respectively, for hydropower generation. For projected future conditions, the proposed operations strategy would increase RRV by 7%, 219%, and 11%, respectively, for water supply and 19%, 136%, and 13% for hydropower generation. Synthesis of the results suggests significant benefits for reliability and resilience of water supply and hydropower are possible with slight operational adjustments. Overall, the integrated performance analysis supports the need to develop an optimized operations rule for Tarbela to adapt to projected climate and demand scenarios.

Keywords: reservoir management; reservoir performance; water management modeling; climate change; water resources planning

1. Introduction Water management has become more complex with human population growth; the emergence of broader performance objectives; and the increase in uncertainty of water availability, energy demand, and flood risk [1–3]. In South Asia, Pakistan is facing grave challenges of increasing population, intensifying precipitation extremes, and broadened expectations for water system performance [4,5]. Once a water-abundant country, Pakistan is now a water-stressed country and is estimated to become water-scarce by 2025, potentially threatening the hydropower and agriculture sector [6,7]. Additionally, floods and droughts pose a significant threat to agriculture and hydropower generation [8,9]. Given Pakistan’s water challenges, large multi-purpose reservoirs such as Tarbela Reservoir constitute a vital component of the country’s water resource infrastructure [10,11]. The Tarbela Reservoir on the Indus River serves multiple purposes, most importantly irrigation supply, hydropower generation, and flood control [12]. The post-Tarbela period has yielded significant economic

Sustainability 2020, 12, 7822; doi:10.3390/su12187822 www.mdpi.com/journal/sustainability Sustainability 2020, 12, x FOR PEER REVIEW 2 of 19 economic benefits for Pakistan by increasing the irrigation command area, enhancing the hydropower capacity, and attenuating flood peaks during the flood season [13]. Although Tarbela Reservoir was designed to serve multiple purposes, the current operations and management practices are biased towards a solitary objective of meeting irrigation supply [14]. SustainabilityThis implies2020 that, 12, 7822hydropower generation and flood control objectives are neglected in the current2 of 18 operations compromising the potential broader performance. According to the Pakistan Water & Power Development Authority (WAPDA), 70% of water discharge is through the spillways benefits(uncontrolled for Pakistan releases), by increasingwhich do thenot irrigationgenerate commandhydropower area, [14]. enhancing This situation the hydropower is alarming capacity, as the andenergy attenuating needs of floodPakistan peaks are during increasing the flood by 7.9% season annually [13]. (currently, Tarbela contributes to 30% of the country’sAlthough hydropower). Tarbela Reservoir Similarly, was designedthese uncont to serverolled multiple releases purposes, (which occur the currentwhen flow operations in the andIndus management is greater than practices 11,326 arem3/s) biased havetowards resulted ain solitary 13 flood objective events causing of meeting significant irrigation damage supply to [ 14the]. Thisagriculture implies sector. that hydropower generation and flood control objectives are neglected in the current operationsTo meet compromising its objectives, the Tarbela potential Reservoir broader is op performance.erated following According a rule tocurve the that Pakistan defines Water the &desired Power water Development surface elevation Authority (and (WAPDA), thus stored 70% water of water volume) discharge of the isreservoir through at thea given spillways time (uncontrolled(Figure 1). Based releases), upon the which operations, do not the generate rule curv hydropowere has three characteristic [14]. This situation components: is alarming (i) falling as thelimb energy (September–May), needs of Pakistan (ii) minimum are increasing level byperiod 7.9% (May–June), annually (currently, and (iii) Tarbelarising limb contributes (July–August). to 30% ofReleases the country’s from the hydropower). reservoir as depicted Similarly, by these the uncontrolledfalling limb and releases minimum (which level occur period when are flow based in theon 3 Indusirrigation is greater demands, than also 11,326 known m / s)as haveKirmani resulted demands in 13 in flood Pakistan, events as causing shown in significant Figure 2 [15]. damage However, to the agriculturethis irrigation sector. demand pattern and the corresponding release pattern from the rule curve present a misleadingTo meet case its objectives,of irrigation Tarbela demands Reservoir being isthe operated highest followingfor June–August. a rule curve Recent that studies defines have the desiredshown waterthat the surface current elevation and future (and crop thus water stored demand water volume) pattern, of represented the reservoir by at crop a given evapotranspiration time (Figure1). Based(ETo), is upon higher the in operations, the months the of ruleApril–June curve has(Tab threele 1) [16,17]. characteristic Moreover, components: the available (i) fallinghead in limb the (September–May),minimum level period (ii) minimum and the risi levelng periodlimb does (May–June), not correspond and (iii) to rising the high limb hydropower (July–August). demand Releases for fromthose themonths. reservoir Figure as 2 depicted shows that by thethe fallinghydropower limb and demand minimum is highest level for period the months are based of June–August, on irrigation demands,but the rule also curve known maintains as Kirmani a low demandsreservoir inlevel Pakistan, in June as and shown July, incompromising Figure2[ 15]. its However, hydropower this irrigationgeneration. demand pattern and the corresponding release pattern from the rule curve present a misleadingPart of case the ofproblem irrigation is demandsthe reliance being of theTarbel highesta Reservoir for June–August. operational Recent planning studies on have the shownsingle thatcriterion the current of irrigation and future supply crop reliability. water demand The use pattern, of a single represented performance by crop criterion evapotranspiration (typically reliability (ETo), isof higherirrigation in thesupply) months is common of April–June for reservoir (Table1 )[operations16,17]. Moreover, around the the world available but is head known in the to be minimum limited levelin describing period and the the strengths rising limb and does weaknesses not correspond of th toe reservoir the high hydropower operations demand[18,19]. Improvement for those months. in Figureoperational2 shows performance that the hydropower is likely feasible demand if decision is highests can for consider the months multiple of June–August, criteria, and in but particular, the rule curvereliability, maintains resilience, a low and reservoir vulnerability level in (RRV) June and [20,21]. July, compromising its hydropower generation.

480

460

440

420 (m) 400

380

360 Reservoir Water Surface amsl Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Months

Figure 1. CurrentCurrent rule curve for the Tarbela Reservoir on a daily time scale.

Table 1. Monthly crop evapotranspiration (ETo) in mm for Northern Pakistan.

Scenario Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Current 68 84 130 173 218 215 194 177 155 124 83 64 1 ◦C Rise 69 85 131 175 219 216 197 178 156 126 84 65 2 ◦C Rise 71 87 133 178 224 221 200 182 160 129 86 67 3 ◦C Rise 76 93 141 186 232 229 207 189 166 134 90 72 Sustainability 2020, 12, 7822 3 of 18 Sustainability 2020, 12, x FOR PEER REVIEW 3 of 19

7,000 3,000

6,000 2,500 5,000

/s) 2,000 3 4,000 1,500 3,000 1,000 Flow (m 2,000 1,000 500 0 0

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec (GWh) Hydropower Demand

Irrigation Demand (m3/s) Hydropower Demand (GWh) Indus River Inflow (m3/s)

Figure 2. Mean monthly inflows, inﬂows, irrigation, and hydropower demand.

Part of the problem is the reliance of Tarbela Reservoir operational planning on the single criterion Table 1. Monthly crop evapotranspiration (ETo) in mm for Northern Pakistan. of irrigation supply reliability. The use of a single performance criterion (typically reliability of irrigationScenario supply) isJan common Feb forMar reservoir Apr operationsMay Jun aroundJul theAug world Sep but Oct is known Nov toDec be limited in describingCurrent the strengths 68 84 and 130 weaknesses 173 218 of the215 reservoir 194 operations177 155 [18124,19 ]. Improvement83 64 in operational1 °C performanceRise 69 is85 likely 131 feasible 175 if decisions 219 216 can 197consider 178 multiple 156 criteria, 126 84and in 65 particular, reliability,2 °C resilience, Rise 71 and 87vulnerability 133 178 (RRV) 224 [20 ,21 221]. 200 182 160 129 86 67 Complicating3 °C Rise matters,76 93 the 141 Tarbela 186 Reservoir 232 is 229 located 207 in a 189 region 166 vulnerable 134 to 90 climate 72 change due to projected variability in precipitation and snowmelt [22–24]. Previous research has indicated Complicating matters, the Tarbela Reservoir is located in a region vulnerable to climate change that an extreme climate change scenario could shift the timing of peak inflows to an earlier month [25]. due to projected variability in precipitation and snowmelt [22,23,24]. Previous research has indicated This implies that the filling of the reservoir (rising limb) should start early to attenuate the flood peak that an extreme climate change scenario could shift the timing of peak inflows to an earlier month and store the flood volume. Furthermore, research suggests that climate change affects the crop sowing [25]. This implies that the filling of the reservoir (rising limb) should start early to attenuate the flood pattern with early sowing, resulting in higher yield [26]. This change in the sowing pattern would peak and store the flood volume. Furthermore, research suggests that climate change affects the crop require early irrigation releases from the reservoir and corresponding low storage volume for the sowing pattern with early sowing, resulting in higher yield [26]. This change in the sowing pattern falling limb. Along with the above mentioned operational stressors, climate change imparts further would require early irrigation releases from the reservoir and corresponding low storage volume for stress on the physical aspects of Tarbela Reservoir due to a low storage coefficient of 0.11 (ratio of the falling limb. Along with the above mentioned operational stressors, climate change imparts live storage to mean annual inflows) and a high draft ratio of 0.97 (ratio of annual demands to the further stress on the physical aspects of Tarbela Reservoir due to a low storage coefficient of 0.11 mean annual inflows) [14]. The low storage coefficient compromises the Tarbela Reservoir’s ability to (ratio of live storage to mean annual inflows) and a high draft ratio of 0.97 (ratio of annual demands regulate extreme hydrological flow variability. to the mean annual inflows) [14]. The low storage coefficient compromises the Tarbela Reservoir’s The efforts to develop an integrated decision support tool for assessing the Indus basin system ability to regulate extreme hydrological flow variability. performance were initiated in 1980s by WAPDA and introduced an Indus Basin Model (IBM) The efforts to develop an integrated decision support tool for assessing the Indus basin system and Indus Basin Model Revised (IBMR) for decision-making in water resource investments [7]. performance were initiated in 1980s by WAPDA and introduced an Indus Basin Model (IBM) and However, the IBM aims to maximize the net economic benefits from agriculture and does not Indus Basin Model Revised (IBMR) for decision-making in water resource investments [7]. However, consider water coverage, reliability, resilience, vulnerability, and demand priorities (agricultural the IBM aims to maximize the net economic benefits from agriculture and does not consider water and flood-control). Furthermore, previous studies on Tarbela have focused mainly on rule curve coverage, reliability, resilience, vulnerability, and demand priorities (agricultural and flood-control). optimization methodologies to improve the performance of the reservoir. For instance, Rashid et al., Furthermore, previous studies on Tarbela have focused mainly on rule curve optimization (2017) used a genetic algorithm for multi-objective optimization to reduce irrigation deficit in Tarbela [27]. methodologies to improve the performance of the reservoir. For instance, Rashid et al. (2017) used a This study proposed a low reservoir level (as compared to the current rule curve) for all months genetic algorithm for multi-objective optimization to reduce irrigation deficit in Tarbela [27]. This exceptstudy proposed August and a low September. reservoir Althoughlevel (as compared this approach to the managed current rule to reduce curve) irrigationfor all months deficit except by a 3 smallAugust amount and September. (< 1.23 km Although), no improvement this approach was managed achieved to for reduce hydropower irrigation generation deficit by and a floodsmall damage reduction. A similar irrigation-centric rule curve optimization approach was employed by amount (< 1.23 Km3), no improvement was achieved for hydropower generation and flood damage Khanreduction. et al., A (2012), similar which irrigation-centric recommended rule further curve reducingoptimization the minimumapproach reservoirwas employed level fromby Khan 417 toet 396al. (2012), m [15 ].which The underlyingrecommended argument further behind reducing this th approache minimum was reservoir that the level rule from curves 417 resulting to 396m from[15]. theThe minimizationunderlying argument of irrigation behind deficits this should approach be preferred, was that since the Tarbela rule wascurves developed resulting to maximizefrom the minimization of irrigation deficits should be preferred, since Tarbela was developed to maximize irrigation benefits. These studies have strived to improve the performance of the Tarbela Reservoir, Sustainability 2020, 12, 7822 4 of 18 irrigation benefits. These studies have strived to improve the performance of the Tarbela Reservoir, yet they have failed to achieve a suitable tradeoff among the multiple objectives of irrigation supply, hydropower generation, and flood control by focusing solely on irrigation benefits. This study was based on a minimalistic approach with a focus towards a generic improvement of the reservoir operation by strategic shifting of the minimum water level in the reservoir rather than developing an optimization algorithm. This approach has been widely used to solve complex problems typical of water resource systems requiring a design strategy based on simulation [21,28,29]. For Tarbela Reservoir, this approach was undertaken after extensive deliberations with all the relevant stakeholders (August 2015 to December 2018) to ensure its analytical simplicity, implementation, relevance to the anticipated impacts of future changes in irrigation demand, hydropower generation, climate, and tradeoffs between operating objectives. The goal of this research was to use an integrated analysis of multiple objectives to prove the feasibility of improving the performance of the Tarbela Reservoir by altering the current rule curve. The reservoir was modeled in a Water Evaluation and Planning (WEAP) model, using a historical 39 years’ streamflow time series for the Indus River along with climate data, irrigation, energy demands and the hydraulic parameters for Tarbela Reservoir. The objective was to explore the effect of alternate operations on the reliability of water supply for irrigation and hydropower and to identify the alternate operation resulting in the highest reliability. Using the proposed rule curve (which resulted in the highest system reliability) and a multicriteria evaluation approach, the performance of the reservoir was evaluated under historical and future conditions. The future conditions were determined using a stress test that explores a range of future possibilities based on historical conditions. Based on preliminary considerations, it was hypothesized that shifting the targeted low storage volumes in the reservoir earlier in the calendar year would lead to improved RRV for the three target objectives of irrigation supply, hydropower generation, and flood control.

1.1. Description of Study Area

1.1.1. Indus River The Indus River emerges from the Tibetan Plateau, enters the north-eastern part of Pakistan, flows through the dry alluvial plains of Punjab and Sindh, and drains into the Arabian Sea. The catchment area of the Indus Basin contributing to the Tarbela Reservoir (the only controlling structure on the Upper Indus River) is 102,435 km2 [30,31]. The mean annual flow of the Indus at Besham Qila, located 80 km upstream of Tarbela Reservoir, is 2336 m3/s [14]. The river flow is highly variable with a standard deviation of 2471 m3/s and recorded low and high flows of 252 m3/s (7 March 2002) and 20,100 m3/s (30 July 2010), respectively [14]. The historical mean monthly inflow shows the typical high intra-annual variation, with about 84% of the annual inflow arriving from May to September, as shown in Figure2[32].

1.1.2. Tarbela Dam and Reservoir

The Tarbela Dam (Latitude: 34◦05023” N, Longitude: 72◦41054” E), located in the Swabi district of Khyber Pakhtunkhwa province, Pakistan, was constructed in 1976 on the Indus River as an outcome of the Indus Waters Treaty (IWT) signed between Pakistan and India in 1960 (Figure3). The IWT had provisions to construct three storage reservoirs, namely Tarbela, Mangla, and Chashma, six barrages, and eight link canals to mitigate the effect of the loss of water due to the restriction on the flows of the rivers in Eastern Pakistan [33]. Besides the primary function of mitigating the effect of loss of water from the eastern rivers and providing a firm base for the development of the irrigation system, a subsidiary goal of hydropower generation was also introduced during the planning phase. The initial design of Tarbela Dam had no provisions for flood control. Flood regulation became an incidental aspect of Tarbela, providing limited flow regulation during the flooding season from July through August. Sustainability 2020, 12, 7822 5 of 18 Sustainability 2020, 12, x FOR PEER REVIEW 5 of 19

FigureFigure 3. 3.Location Location map map of of Tarbela Tarbela Dam Dam and and Reservoir. Reservoir.

1.1.3.1.1.3. Current Current Reservoir Reservoir Operations Operations of of Tarbela Tarbela TarbelaTarbela is is a a large large multipurpose multipurpose reservoir, reservoir, the the operational operational rules rules of whichof which are are now now set accordingset according to theto irrigationthe irrigation requirements requirements of the of provinces. the provinces. From 1976 From to 1993,1976 theto 1993, lack of the a water-sharing lack of a water-sharing agreement betweenagreement the provincesbetween the led provinces to ad hoc operationsled to ad hoc of the operations Tarbela Reservoir. of the Tarbela In 1991, Reservoir. the provinces In 1991, signed the aprovinces formal accord, signed which a formal resulted accord, in which the creation resulted of thein the Indus creation River of System the Indus Authority River System (IRSA) Authority in 1993. The(IRSA) IRSA in is 1993. now The the designatedIRSA is now body the responsibledesignated body for water responsible allocation for among water allocation the provinces among based the onprovinces agreed indicatorsbased on ofagreed the provincial indicators allocations, of the provincial which are allocations, dependent which on the are average dependent river flows.on the Theaverage indents river (water flows. requests) The indents submitted (water byrequests) the provinces submitted form by thethe basisprovinces of the form outflows the basis from of the the reservoir,outflows whichfrom overthe thereservoir, operational which period over of the Tarbela operational resulted inperiod the development of Tarbela ofresulted the rule curve.in the Thedevelopment rule curve of developed the rule bycurve. WAPDA The rule (Figure curve1) shows developed that theby reservoirWAPDA should(Figure be 1) lowered shows that to the the minimumreservoir should level of be 396 lowered m above to the mean minimum sea level level (amsl) of 396m by the above middle mean of sea May level until (amsl) the by 20th the of middle June toof maximize May until releasesthe 20th of for June irrigation to maximize demand. releases The for maximum irrigation level demand. of 472 Th me maximum amsl is the level target of 472m for mid-Augustamsl is the aftertarget attenuating for mid-August the high after flood attenuating peaks in Julythe high and earlyflood August.peaks in The July filling and early should August. be at anThe average filling rateshould of 2.8 be mat/ dayan average up to 460 rate m of amsl 2.8 m/day and 0.3 up m /today 460m from amsl 460 and to 472 0.3 m m/day amsl. from The filling460 to rates472m wereamsl. established The filling by rates considering were established the structural by considering integrity and the safetystructural of Tarbela integrity Dam. and safety of Tarbela Dam.

2. Materials and Methods Sustainability 2020, 12, 7822 6 of 18

Sustainability 2020, 12, x FOR PEER REVIEW 6 of 19 2. Materials and Methods This study was divided into three parts (Figure 4). First, a computer model was created using This study was divided into three parts (Figure4). First, a computer model was created using the Water Evaluation and Planning (WEAP) system. This model was verified using observed data. the Water Evaluation and Planning (WEAP) system. This model was verified using observed data. Second, the WEAP model was used to iterate through an analysis of the effects of altering the current Second, the WEAP model was used to iterate through an analysis of the effects of altering the rule curve by shifting the minimum reservoir level to an earlier period and using irrigation supply current rule curve by shifting the minimum reservoir level to an earlier period and using irrigation and hydropower reliability as initial criteria for performance evaluation. The outcome of this analysis supply and hydropower reliability as initial criteria for performance evaluation. The outcome of this was a proposed rule curve, which resulted in the highest reliability for irrigation supply and analysis was a proposed rule curve, which resulted in the highest reliability for irrigation supply and hydropower generation. The third part of the research was to test the proposed rule curve under hydropower generation. The third part of the research was to test the proposed rule curve under historical and future boundary conditions of inflow and demand based on the RRV criteria for the historical and future boundary conditions of inflow and demand based on the RRV criteria for the three objectives. The WEAP model was applied again to test the proposed rule curve. Two WEAP three objectives. The WEAP model was applied again to test the proposed rule curve. Two WEAP models were developed, one with the current operations rule (CR) and the other with the proposed models were developed, one with the current operations rule (CR) and the other with the proposed operations rule (PR). Each model was analyzed for (i) a historical period (June 1976 to December operations rule (PR). Each model was analyzed for (i) a historical period (June 1976 to December 2013) with historical conditions of climate (inflow) and irrigation demand and (ii) for a future period 2013) with historical conditions of climate (inflow) and irrigation demand and (ii) for a future period with projected climate and demand conditions, based on a stress analysis of historical conditions, with projected climate and demand conditions, based on a stress analysis of historical conditions, where the conditions varied from an increase of 90% to decrease of 90% of the historical values [34,35]. where the conditions varied from an increase of 90% to decrease of 90% of the historical values [34,35]. The evaluation of the proposed rule curve was based on performance improvement in terms of RRV The evaluation of the proposed rule curve was based on performance improvement in terms of RRV under the proposed operations compared to current operations for both historical and future under the proposed operations compared to current operations for both historical and future conditions. conditions.

Figure 4. Flow chart representing overview of methodology. Figure 4. Flow chart representing overview of methodology. 2.1. Water Evaluation and Planning (WEAP) System Model 2.1. Water Evaluation and Planning (WEAP) System Model WEAP is a decision support system that integrates and simulates surface and groundwater resources. It isWEAP built on is water a decision balance support principles system and that can accessintegrates supply and andsimulates demand surface conditions and groundwater with various resources.alternative It scenarios is built on such water as climate,balance principles land use, infrastructure, and can access and supply water and management demand conditions priorities with [36]. variousFor the presentalternative study, scenarios a representative such as model climate, was land created use, for infrastructure, Tarbela Dam and and Reservoir water management using WEAP. priorities [36]. For the present study, a representative model was created for Tarbela Dam and Reservoir using WEAP.

2.2. Data Collection and Input The daily inflow, demand, and physical data for reservoir inflow and release measurements were obtained from WAPDA, the operating authority of the Tarbela Reservoir [14]. The data range Sustainability 2020, 12, 7822 7 of 18

2.2. Data Collection and Input Sustainability 2020, 12, x FOR PEER REVIEW 7 of 19 The daily inflow, demand, and physical data for reservoir inflow and release measurements werecovers obtained the period from from WAPDA, June 1976 the to operating December authority 2013. The of quality the Tarbela of data Reservoir acquired [14 from]. The WAPDA data range was coversdivided the into period three from categories: June 1976 verified, to December acceptable, 2013. an Thed uncertain. quality of The data verified acquired data from refer WAPDA to recorded, was divideddocumented, into three and categories:verified data verified, managed acceptable, by WAPDA. anduncertain. The acceptable The verified quality dataof data refer was to recorded,recorded documented,and documented and verifiedbut not verified, data managed and the by uncertain WAPDA. qu Theality acceptable of data refers quality to of data data not was documented recorded and or documentedverified by WAPDA. but not verified, As the WEAP and the model uncertain of Tarbela quality is of demand-driven, data refers to data it was not documentedvery important or verified to have byaccurate WAPDA. inflow, As the demand, WEAP and model physical of Tarbela data is of demand-driven, the reservoir [37]. it was The very daily important inflow and to havedemand accurate data inflow,were of demand, verified andquality. physical The physical data of thedata reservoir for the reservoir [37]. The were daily mainly inflow andof verified demand quality, data were except of verifiedevaporation, quality. which The physicalwas of dataacceptable for the reservoirquality. wereTarbela mainly Dam of verifiedand Reservoir quality, exceptwere modeled evaporation, by whichdescribing was the of acceptable physical, quality.operation, Tarbela hydropower, Dam and cost, Reservoir and priority were modeled in WEAP. by describingThe WEAP the model physical, was operation,calibrated hydropower,based on the cost,comparison and priority of the in WEAP.observed The releases WEAP to model the wasWEAP calibrated model releases. based on The the comparisoncalibration period of the observedfrom June releases 1976 to toDecember the WEAP 2001 model was releases. selected Theas a calibrationrepresentative period record from of June the 1976entire to 1976 December to 2013 2001simulation was selected period, as considering a representative mean recordannual offlow the and entire extreme 1976 toevents 2013 (Figure simulation 5a). period,Nash–Sutcliffe considering efficiency mean annual(NSE) was flow used and extremeto calcul eventsate the (Figureaccuracy5a). of Nash–Sutcli the observedffe data efficiency and model (NSE) wasdata used[38]. A to NSE calculate value the close accuracy to 1 shows of the that observed the accu dataracy of and the model model data is high. [38]. The A time NSE series value of close the toreleases 1 shows from that the the model accuracy with of the the increase model isin high.the evaporation The time seriesby 30% of (0.026 the releases ft/day) from resulted the modelin the withhighest the NSE increase index in of the 0.77. evaporation Therefore, by a 30% 30% increase (0.026 ft /inday) evaporation resulted inwas the selected highest as NSE the indexchosen of value 0.77. Therefore,for calibration. a 30% The increase timing in evaporationof the model was peaks selected matched as the well chosen with value the observed for calibration. discharge The values. timing ofHowever, the model the peaks magnitude matched of well the withpeak the values observed was higher discharge than values. the observed However, discharge the magnitude values. of The the peakvalidation values period was higher was selected than the as observedJanuary 2002 discharge to December values. 2013. The The validation validation period results was gave selected an NSE as Januaryindex value 2002 of to 0.74 December (Figure 2013. 5b). The validation results gave an NSE index value of 0.74 (Figure5b).

500,000 800,000 450,000 700,000 400,000 350,000 600,000 300,000 500,000 250,000 400,000 200,000

Releases (cfs) Releases 300,000

150,000 Release (cfs) 100,000 200,000 50,000 100,000 0 0

WEAP Model Discharge Observed Discharge WEAP Model Discharge Observed Discharge

(a) (b)

FigureFigure 5.5. HydrographsHydrographs ofof thethe observedobserved andand simulatedsimulated releasesreleases showingshowing ((aa)) calibratedcalibrated modelmodel ﬁtfit andand ((bb)) validationvalidation period.period.

2.3.2.3. Exploration of Altered ReservoirReservoir OperationsOperations TheThe alteredaltered operationsoperations ofof TarbelaTarbela DamDam werewere developeddeveloped byby changingchanging thethe targettarget waterwater surfacesurface elevationselevations inin thethe reservoirreservoir (i.e.,(i.e., thethe rulerule curve)curve) withwith thethe intentintent ofof improvingimproving thethe performanceperformance underunder historicalhistorical andand futurefuture conditions.conditions. The altered operationsoperations to explore were guidedguided byby thethe followingfollowing observationsobservations ofof currentcurrent operations:operations: (1)(1) Under current operations, the water level in the reservoirreservoir isis maintainedmaintained atat thethe minimumminimum fromfrom mid-May to to mid-June. mid-June. By By maintaining maintaining the the reserv reservoiroir water water surface surface level level at the at theminimum, minimum, the waterthe water releases releases are maximized, are maximized, but at but the at expe the expensense of hydropower of hydropower gene generationration potential. potential. The releasesThe releases should should also also be maximized be maximized in inearlier earlier months months to to better better complement complement irrigation irrigation demands. demands. Therefore, one one proposed proposed operation operation change change is isto toadjust adjust the thetiming timing of the of minimum the minimum reservoir reservoir level bylevel shifting by shifting it to itan to earlier an earlier period. period. The The rule rule curve curve can can thus thus be be modified modiﬁed but but adhere adhere to threethree constraints. First,First, thethe maximummaximum reservoir reservoir level level should should be targetedbe targeted for latefor late August August because because of high of high flows during July and August. Second, the proposed earlier drawdown of the reservoir must follow the water indent requirements from the provinces which are based on crop water demand. Third, drawdown and filling rate should not exceed 3 m/day to ensure the structural integrity of the dam. Figure 6 shows the resulting set of rule curves with varying minimum Sustainability 2020, 12, 7822 8 of 18

flows during July and August. Second, the proposed earlier drawdown of the reservoir must follow the water indent requirements from the provinces which are based on crop water demand. SustainabilityThird, 2020 drawdown, 12, x FOR andPEERfilling REVIEW rate should not exceed 3 m/day to ensure the structural integrity8 of of19 the dam. Figure6 shows the resulting set of rule curves with varying minimum reservoir levels. reservoirUsing WEAP levels. simulations, Using WEAP the simulations, recommended the rule recommended curve (herein rule called curve proposed (herein called operations) proposed can operations)be set based can on thebe bestset tradeobasedff onfor reliabilitythe best trad betweeneoff irrigationfor reliability and hydropowerbetween irrigation objectives. and (2) hydropowerSince flood control objectives. is not addressed by the proposed changes in the timing of the minimum water (2) Sincelevel orflood other control changes is innot the addressed rule curve, by structural the proposed changes changes to upgrade in the the timing capacity of the of theminimum tunnels water(increasing level or maximum other changes hydraulic in the outflow) rule curve, were st investigatedructural changes as a solution.to upgrade the capacity of the tunnels (increasing maximum hydraulic outflow) were investigated as a solution.

480 460 440 420 400

Elevation amsl (m) 380 360 Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Jan-Feb Feb-March March-April April-May May-June

Figure 6. RuleRule curves curves with with different diﬀerent peri periodsods of minimum reservoir levels.levels.

2.4. Performance Performance Evaluation of Proposed Operations Following Hashimoto et al.,al. (1982), (1982), the the performanc performancee evaluation evaluation (RRV (RRV of of ea eachch objective) of the proposed operations was based on comparing currentcurrent operations and proposedproposed operations under historical and future climate (inflow)(inflow) and irrigationirrigation demand:demand: current operations (CR) vs. proposed operations (PR) under historical climate and demand conditions (H) (H-CR vs. H-PR) and current operations vs.vs. proposedproposed operationsoperations under under future future climate climate and and demand demand conditions conditions (F) (F-CR(F) (F-CR vs. F-PR)vs. F-PR) [39]. The[39]. futureThe future conditions conditions refer refer to projected to projected inflow inflow and irrigation and irrigation water water demand demand conditions conditions based based upon uponthe stress the stress test on test historical on historical conditions. conditions. The period The period for the for future the future conditions conditions was fromwas from January January 2014 2014to December to December 2050. 2050. The The number number of yearsof years in in the the future future was was the the same same as as that that used used forfor thethe historical conditions because the future conditions were not precisely calculated but assumed as varying ranges between -90%90% to to 90% 90% (10% (10% interval) interval) of the of thehistorical historical conditions. conditions. Based Based on previous on previous studies, studies, this range this − ofrange future of futureconditions conditions is plausible is plausible to represent to represent the extreme the extreme conditions, conditions, i.e., extensive i.e., extensive floods floods and long- and termlong-term droughts droughts and very and high very highand low and irrigation low irrigation water water demands demands [40]. [40]. Three performance indices,indices, viz.viz. reliability, reliability, resilience, resilience, and and vulnerability, vulnerability, were were used used for performance evaluation. Reliability is defined defined as the probability that the system status status remains remains in in a a satisfactory satisfactory state. state. Water forfor irrigationirrigation supply and hydropower generation reliability for each scenario were calculated internally in WEAP. WEAPWEAP calculatescalculates reliabilityreliability as:as:

𝑁𝑢𝑚𝑏𝑒𝑟Number o f time 𝑜𝑓 periods 𝑡𝑖𝑚𝑒 o f zero 𝑝𝑒𝑟𝑖𝑜𝑑𝑠 shortage(Zt) 𝑜𝑓𝑧𝑒𝑟𝑜𝑠ℎ𝑜𝑟𝑡𝑎𝑔𝑒(𝑍𝑡) 𝑅𝑒𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑦Reliability (𝑅𝑡(Rt))== (1)(1) 𝑇𝑜𝑡𝑎𝑙Total Time Period 𝑇𝑖𝑚𝑒(Tp 𝑃𝑒𝑟𝑖𝑜𝑑(𝑇𝑝) ) where, 0 ≤ Rt ≤ 100, Zt ≤ Tp. where, 0 Rt 100, Zt Tp. For water≤ ≤ supply and≤ hydropower reliability, the time period of zero shortage means a sufficient For water supply and hydropower reliability, the time period of zero shortage means a suﬃcient volume of water was supplied to the demand node and the turbines for hydropower generation. The volume of water was supplied to the demand node and the turbines for hydropower generation. flood control reliability was externally calculated based on the same definition of reliability as used for water supply and hydropower. For flood control reliability, the time period of zero shortage refers to the days when the releases from the reservoir were less than 11,326 m3/s. Resilience describes how quickly a system recovers from the failure state after a failure occurs. The irrigation supply and hydropower generation resilience were calculated externally by exporting the supply delivered and hydropower generation data from the Result view in WEAP. The actual Sustainability 2020, 12, 7822 9 of 18

The flood control reliability was externally calculated based on the same definition of reliability as used for water supply and hydropower. For flood control reliability, the time period of zero shortage refers to the days when the releases from the reservoir were less than 11,326 m3/s. Resilience describes how quickly a system recovers from the failure state after a failure occurs. The irrigation supply and hydropower generation resilience were calculated externally by exporting the supply delivered and hydropower generation data from the Result view in WEAP. The actual irrigation demand and hydropower demand and the exported Excel sheets of supply delivered and hydropower generation were used to determine the number of times the system follows a failure state (F) from the satisfactory state (S) and a number of days the system was in the failure state (Tf). The failure state was defined as when the required volume of water could not be released from the reservoir for irrigation demand or hydropower generation. The frequency of a satisfactory state following an unsatisfactory state is the basis of resilience:

Number o f times a satis f actory state f ollows an unsatis f actory state(S F) Resilience = → (2) Number o f unsatis f actory value occurs

The flood control resiliency was calculated externally by exporting the reservoir release data from WEAP. The number of times a satisfactory state (release less than 11,326 m3/s) followed an unsatisfactory state (number of times when the flow exceeded 11,326 m3/s) and the number of days’ releases exceeding the threshold values were calculated and used in the above equation. Vulnerability refers to the likelihood of failure. The vulnerability of irrigation supply, hydropower generation, and flood control was calculated externally by using the following equation:

Sum o f Positive values o f (target delivery) Vulnerability = − (3) Number o f times an unsatis f actory value occurs

The performance evaluation for the entire period with proposed operations over the current operations was calculated as the percentage improvement in the system performance for each objective with proposed operations. Percentage improvement in performance with proposed operations for any objective was computed based on Equation (4).

( H PR0 Rel) ( H CR0 Rel) % improvement = 0 − − 0 − 100 (4) H PR ∗ − Positive values show improvement, while negative values show a decrease in the performance indices of the objective. The above equation can be substituted for future scenarios (F-CR vs. F-PR) and performance measures (resiliency and vulnerability).

3. Results and Discussion

3.1. Evaluation of Altered Reservoir Operations The simulations of the WEAP model with altered rule curves under the historical conditions showed that the reliability of hydropower and irrigation supply was sensitive to the changes in the minimum reservoir level (Figure7). The best combination of hydropower generation and meeting irrigation demand reliability occurred when the minimum reservoir level was set during April through May (Figure7). This was because the minimum reservoir level in April–May increased the releases and complemented the irrigation demand pattern. Furthermore, early ﬁlling (after mid-May) resulted in more available head during the months (June–August) having the highest hydropower demand. Thus, a rule curve presented in Figure8 was proposed for the integrated performance of the Tarbela Reservoir. Sustainability 2020, 12, 7822 10 of 18 Sustainability 2020, 12, x FOR PEER REVIEW 10 of 19

60 50 40 30 20 Irrigation supply

Reliability (%) Reliability 10 0 Hydropower Reliability

Months with Minimum Reservoir Level

Figure 7. ReliabilityReliability of of irrigation irrigation supply supply and and hydropower generation by altering the minimum Sustainability 2020, 12, x FOR PEER REVIEW 11 of 19 reservoir level.level.

480

460

440

420

400 Elevation amsl amsl (m) Elevation

380

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

Proposed Rule Curve Current Rule Curve

FigureFigure 8. 8.ProposedProposed rule curvecurve of of Tarbela Tarb showingela showing two distincttwo distinct regions regions of irrigation of irrigation supply improvement supply improvement(inclined dashes) (inclined and dashes) hydropower and hydropower generation improvementgeneration im (horizontalprovement lines).(horizontal lines). The improvement in flood control reliability required structural changes in the outflow capacity The improvement in flood control reliability required structural changes in the outflow capacity of the dam, as the altered operations could not substantially address flooding downstream of Tarbela of the dam, as the altered operations could not substantially address flooding downstream of Tarbela because of the low storage coefficient. The analysis showed that the lowest outlet capacity resulted in because of the low storage coefficient. The analysis showed that the lowest outlet capacity resulted the lowest of flood control reliability and vice versa. The flood control reliability was maximum when in the lowest of flood control reliability and vice versa. The flood control reliability was maximum the release capacity from the reservoir was 11,326 m3/s (Figure9). when the release capacity from the reservoir was 11,326 m3/s (Figure 9). Although the difference in the percentage of flood control reliability is a fraction of a percent, this

small value represents a ﬂood occurring and causing millions of dollars’ worth of damage and lost economic productivity [35]. Sustainability 2020, 12, 7822 11 of 18 Sustainability 2020, 12, x FOR PEER REVIEW 12 of 19

100

99.9

99.8

99.7 Reliability (%) Reliability 99.6

99.5 1500 3500 5500 7500 9500 11500 Maximum Hydraulic Outflow (m3/s)

Figure 9. PercentagePercentage change change in in flood ﬂood control relia reliabilitybility with changes in hydraulic ﬂow.flow.

3.2. PerformanceAlthough the Improvement difference with in the Proposed percentage Operations of floo ford Historicalcontrol reliability and Future is Conditionsa fraction of a percent, this small value represents a flood occurring and causing millions of dollars’ worth of damage and 3.2.1. Historical Conditions (H-CR vs. H-PR) lost economic productivity [35]. The measure of improvement (RRV) for the three objectives for the period of the analysis is presented3.2. Performance as the Improvement cumulative with distribution Proposed Operations function forfor Historical each performance and Future measureConditions (Figure 10a,b). The likelihood of occurrence of higher reliability and resilience increases towards the right and when the3.2.1. curve Historical follows Conditions a narrow band(H-CR as vs. with H-PR) PR. Based on the cumulative distribution frequency (CDF) curves, the improvement in the system performance is summarized as a 3 3 matrix with three The measure of improvement (RRV) for the three objectives for the period× of the analysis is objectivespresented andas the three cumulative performance distribution measures, function as shown for each in Table performance2. measure (Figure 10a,b). The likelihood of occurrence of higher reliability and resilience increases towards the right and when the curveTable follows 2. Percentage a narrow increase band inas reliability, with PR. resilience, Based on and the vulnerability cumulative (RRV) distri withbution proposed frequency operations (CDF) for historical conditions. curves, the improvement in the system performance is summarized as a 3 × 3 matrix with three objectives and three performanceObjectives measures, Reliabilityas shown in Table Resilience 2. Vulnerability Water Supply 17 * 67 * 7 Table 2. PercentageHydropower increase Generation in reliability, resilienc 34 *e, and 346 vulnerability * 22(RRV) * with proposed operations for historicalFlood conditions. Control 0.3 * - - * StatisticallyObjectives significant difference at 95%Reliability confidence level based onResilience two sample t-test usingVulnerability proposed rule curve of Tarbela compared to the current rule curve. Water Supply 17 * 67 * 7 Hydropower The water supply and hydropower34 objectives* show improvement346 * in terms22 of * reliability and Generation resilience, but this improvement renders the system more vulnerable. The tradeoffs between the RRV Flood Control 0.3 * - - criteria have been analyzed by past researchers [18,28,41]. It was demonstrated that if the system * Statistically significant difference at 95% confidence level based on two sample t-test using proposed reliability is increased and/or maximum length of consecutive deficits decrease, the vulnerability of rule curve of Tarbela compared to the current rule curve. the system to large deficits increases. For Tarbela Reservoir, the reduced uncertainty and increased reliability and resilience for irrigation supply and hydropower were achieved by restricting releases to The water supply and hydropower objectives show improvement in terms of reliability and the outlet structures (after passing through the hydropower turbines) instead of spillways. In addition, resilience, but this improvement renders the system more vulnerable. The tradeoffs between the RRV the raising of the water level during the wet season and shifting of the minimum reservoir level to criteria have been analyzed by past researchers [18,28,41]. It was demonstrated that if the system April–May increased the volume of water available for hydropower generation and irrigation water reliability is increased and/or maximum length of consecutive deficits decrease, the vulnerability of supply, respectively. The flood control objective shows a minor improvement in reliability. Because of the system to large deficits increases. For Tarbela Reservoir, the reduced uncertainty and increased the lack of failure in terms of flood, resilience and vulnerability could not be calculated. reliability and resilience for irrigation supply and hydropower were achieved by restricting releases to the outlet structures (after passing through the hydropower turbines) instead of spillways. In addition, the raising of the water level during the wet season and shifting of the minimum reservoir level to April–May increased the volume of water available for hydropower generation and irrigation water supply, respectively. The flood control objective shows a minor improvement in reliability. Because of the lack of failure in terms of flood, resilience and vulnerability could not be calculated.

Sustainability 2020, 12, 7822 12 of 18 Sustainability 2020, 12, x FOR PEER REVIEW 13 of 19

(a)

(b)

FigureFigure 10. TheThe plot plot of of cumulative cumulative distribution distribution frequency frequency (CDF) (CDF) of of objectives objectives ( (aa)) irrigation irrigation supply supply resilienceresilience and and ( (bb)) flood ﬂood control. control.

3.2.2. Future Conditions (F-CR vs. F-PR) The system response to future demand and climate conditions with current and proposed operations was analyzed by developing contour graphs for each objective and performance measure (Figure 11a,b and Figure 12a,b). The contour graphs were developed with the ranges of climate-induced inﬂow changes on the y-axis and demand changes on the x-axis. They provide a framework for determiningSustainability the 2020 system, 12, x FOR response PEER REVIEW under a speciﬁc future demand and climate condition. 15 of 19

(a)

(b)

FigureFigure 11. Contour 11. Contour graphs graphs of resilience of resilience for the for irrigation the irrigation supply supply objective objective (a) proposed, (a) proposed, (b) current (b) current operations. operations. Sustainability 2020, 12, 7822 14 of 18 Sustainability 2020, 12, x FOR PEER REVIEW 16 of 19

(a)

(b)

Figure 12. ContourContour graphs graphs of ofresilience resilience for forthe thehydropower hydropower generation generation objective objective (a) proposed, (a) proposed, (b) (currentb) current operations. operations.

4. ConclusionsThe three objectives and Recommendations of water supply, hydropower generation, and flood control exhibited distinct behavior in terms of their respective sensitivities to inflow and demand variations. The RRV for In this study, a WEAP model for the Tarbela Reservoir was created and used to analyze irrigation water supply was very sensitive to the changes in inflow and downstream demands. integrated performance for historical and future scenarios against multiple objectives and multiple The response for irrigation supply can be attributed to the fact that meeting downstream demand criteria. The limitations of the current operations of Tarbela Dam were explored, and a possible new was dependent and coupled to the variation in inflows and inflow–demand relationship. On the operations approach was identified. The methodology of quantifying the improvement under other hand, the RRV for hydropower generation and flood control was not found to be sensitive proposed operations was based on computing the RRV of each objective (irrigation demand, to the changes in the downstream irrigation demand for a given inflow condition. This is because hydropower generation, flood control) under the current operation strategy and the proposed altered operation strategy, for both historical and future climate and demand conditions. The RRV of Sustainability 2020, 12, 7822 15 of 18 hydropower generation and flood control require constraining the water and altering the releases from the dam. Storage of water in the reservoir is dependent on the energy requirement, without any linkage to the downstream demands, and sensitive to the variation in the inflows only [42–44]. In terms of performance, the system showed the highest improvement in the resilience metric as shown in Table3.

Table 3. Percentage increase in RRV with proposed operations for future conditions.

Objectives Reliability Resilience Vulnerability Water Supply 7 219 * 11 Hydropower Generation 19 * 136 * 13 Flood Control 2 * 33 * 39 * − − * Statistically significant difference at 95% confidence level based on two-sample t-test.

The contour maps developed for the future conditions provide a range of system responses in terms of RRV from which system response for any plausible future scenarioasin (UIB) from 15% − can be determined. For instance, Lutz et al., (2014) suggested a range of future inﬂow variability in the upper Indus b to 60%, with respect to historical time series from 1971–2000, by the end of 21st century [22]. Similarly, Amin et al., (2018) determined future increase in irrigation demand by 50% in the UIB in 2050 compared to 2010 [45]. From Figure 11, decreases in inﬂow by 15% and increase in demand by 50% shows that the irrigation supply resilience will be around 14% for PR and 4% for CR. The contour maps are useful to examine system response across a range of what-if conditions.

4. Conclusions and Recommendations In this study, a WEAP model for the Tarbela Reservoir was created and used to analyze integrated performance for historical and future scenarios against multiple objectives and multiple criteria. The limitations of the current operations of Tarbela Dam were explored, and a possible new operations approach was identified. The methodology of quantifying the improvement under proposed operations was based on computing the RRV of each objective (irrigation demand, hydropower generation, flood control) under the current operation strategy and the proposed altered operation strategy, for both historical and future climate and demand conditions. The RRV of respective operations and conditions combinations were computed and examined to determine if improvements could be achieved. The main conclusions from the research are as follows: (1) Effects of rule curve change on reservoir performance: The timing of the minimum target water elevation in Tarbela Reservoir was shifted one month earlier in the calendar year to April–May. The shifting of the minimum reservoir level increased the performance of Tarbela Reservoir for both irrigation supply and hydropower objectives. This shifting of the rule curve will ensure an increase in sustainability and resilience to future uncertainties of climate and irrigation and hydropower demand. (2) Maximum hydraulic outflow: The maximum hydraulic outflow played a pivotal role in controlling the flooding downstream of Tarbela. It was concluded that the maximum release capacity from the tunnels needs to be increased to reduce spillway discharges that do not generate hydropower. The downstream channel of Tarbela can safely convey 11,326 m3/s, which was tested and shown to lead to improved performance. (3) Percentage improvement in RRV with proposed operations: The performance of the reservoir operations with proposed operations over current operations was evaluated for the entire period for historical and future climate and demand conditions. The analysis resulted in the exploration of gains in objectives with the proposed operations. The performance evaluations of water supply for irrigation and hydropower objectives showed improvement in reliability and resilience with an increase in vulnerability. The flood control objective resulted in an increase in reliability and a decrease in resilience and vulnerability. Sustainability 2020, 12, 7822 16 of 18

5. Future Work and Limitations of Study The limitations with the existing model are that the future climate and demand conditions were not forecasted. They were based on arbitrary percentage increases and decreases of the historical climate and inflow observations. Although simplified, the approach does provide a reasonable stress test of the system that is useful for evaluating proposed changes under uncertain moderate and extreme conditions, giving confidence to decision making. The improvement in the performance measures was only evaluated as the percentage increase or decrease of meeting objectives compared with current operations. This approach does not consider the tangible benefits such as cost/profit per unit release of water from the reservoir. This should be investigated as part of future work to confirm the conclusions drawn from determining the performance improvements. Another uncertain factor not considered in this study was the change in live storage volume due to sedimentation and possible future upstream dam construction (especially Diamir Bhasha Dam). These factors could impact the storage volume required to meet the objectives, which could affect the system performance in terms of RRV. Additional alterations in operations may be needed. Finally, the approach of using simulation to investigate selected operation alterations could be refined to include optimization. The investigation was thorough; therefore, optimization is not expected to significantly alter the results. However, as the solution space and constraints are made more complicated, optimization may be needed.

Author Contributions: Conceptualization and methodology, A.R., and S.B.; resources and data acquisition, R.B.; and D.H.; writing—original draft preparation, A.R.; writing—review and editing, A.R., S.B., and D.H.; visualization, A.R., S.B., R.B., and D.H.; supervision, S.B. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the United States Agency for International Development through a U.S,-Pakistan Center for Advanced Studies in Water (University of Utah). Acknowledgments: The authors are sincerely grateful to WAPDA for data and technical help and two anonymous reviewers for their suggestions to improve the quality of manuscript. Conflicts of Interest: The authors declare no conflict of interest. Disclaimer: This study was made possible by the support of the United States Government and the American people through the United States Agency for International Development (USAID). The contents are the sole responsibility of the authors and do not necessarily reflect the views of USAID or the United States Government.

References

1. Pereira, L.S.; Oweis, T.; Zairi, A. Irrigation management under water scarcity. Agric. Water Manag. 2002, 57, 175–206. [CrossRef] 2. Somlyódy, L.; Varis, O. Freshwater under pressure. Int. Rev. Environ. Str. 2006, 6, 181–204. 3. Wheater, H.; Evans, E. Land use, water management and future flood risk. Land Use Policy 2009, 26, S251–S264. [CrossRef] 4. Kijne, J.W. Lessons learned from the change from supply to demand water management in irrigated agriculture: A case study from Pakistan. Water Policy 2001, 3, 109–123. [CrossRef] 5. Qureshi, A.S.; McCornick, P.G.; Sarwar, A.; Sharma, B.R. Challenges and prospects of sustainable groundwater management in the Indus Basin, Pakistan. Water Resour. Manag. 2010, 24, 1551–1569. [CrossRef] 6. Arfan, M.; Lund, J.; Hassan, D.; Saleem, M.; Ahmed, A. Erratum: Assessment of spatial and temporal flow variability of the Indus River. Resources 2019, 8, 103. [CrossRef] 7. Hassan, D.; Rais, M.N.; Ahmed, W.; Bano, R.; Burian, S.J.; Ijaz, M.W.; Bhatti, F.A. Future water demand modeling using water evaluation and planning: A case study of the Indus basin in Pakistan. Sustain. Water Resour. Manag 2019, 5, 1903–1915. [CrossRef] 8. Kirsch, T.D.; Wadhwani, C.; Sauer, L.; Doocy, S.; Catlett, C. Impact of the 2010 Pakistan floods on rural and urban populations at six months. PLoS Curr. 2012, 4, 4fdfb212d2432. [CrossRef] 9. Looney, R. Economic impacts of the floods in Pakistan. Contemp. South Asia 2012, 20, 225–241. [CrossRef] Sustainability 2020, 12, 7822 17 of 18

10. Tate, E.L.; Farquharson, F.A. Simulating reservoir management under the threat of sedimentation: The case of Tarbela dam on the river Indus. Water Resour. Manag. 2000, 14, 191–208. [CrossRef] 11. Khan, N.M.; Tingsanchali, T. Optimization and simulation of reservoir operation with sediment evacuation: A case study of the Tarbela Dam, Pakistan. Hydrol. Process. Int. J. 2009, 23, 730–747. [CrossRef] 12. Tarar, R.N. Performance of Tarbela Dam project. In In Proceedings of Pakistan Engineering Congress, 69th Annual Session Proceedings Report; Punjab, Pakistan, 2006; pp. 8–10. Available online: https://www.pecongress.org. pk/images/upload/books/632.pdf (accessed on 30 August 2020). 13. Water & Power Development Authority. Available online: http://www.wapda.gov.pk/index.php/investor-s- corner/annual-report (accessed on 24 April 2018). 14. The World Commission on Dams. Available online: https://energypedia.info/wiki/World_Commission_on_ Dams_(WCD)_Report (accessed on 27 April 2018). 15. Khan, N.M.; Babel, M.S.; Tingsanchali, T.; Clemente, R.S.; Luong, H.T. Reservoir optimization-simulation with a sediment evacuation model to minimize irrigation deficits. Water Resour. Manag. 2012, 26, 3173–3193. [CrossRef] 16. Naheed, G.; Mahmood, A. Water requirement of wheat crop in Pakistan. Pak. J. Meteorol. 2009, 6, 89–97. 17. Naheed, G.; Rasul, G. Projections of crop water requirement in Pakistan under global warming. Pak. J. Meteorol. 2010, 7, 45–51. 18. Moy, W.S.; Cohon, J.L.; Revelle, C.S. A programming-model for analysis of the reliability, resilience, and vulnerability of a water-supply reservoir. Water Resour. Res. 1986, 22, 489–498. [CrossRef] 19. Vogel, R.M.; Bolognese, R.A. Storage-reliability-resilience-yield relations for over-year water-supply systems. Water Resour. Res. 1995, 31, 645–654. [CrossRef] 20. Jain, S.K.; Bhunya, P.K. Reliability, resilience and vulnerability of a multipurpose storage reservoir/confiance, résilience et vulnérabilité dun barrage multi-objectifs. Hydrol. Sci. J. 2008, 53, 434–447. [CrossRef] 21. Mateus, M.C.; Tullos, D. Reliability, sensitivity, and vulnerability of reservoir operations under climate change. J. Water Resour. Plan. Manag. 2017, 143, 04016085. [CrossRef] 22. Lutz, A.F.; Immerzeel, W.W.; Shrestha, A.B.; Bierkens, M.F.P. Consistent increase in high Asia’s runoff due to increasing glacier melt and precipitation. Nat. Clim. Chang. 2014, 4, 587–592. [CrossRef] 23. Wang, W.; Xiang, Y.; Gao, Y.; Lu, A.; Yao, T. Rapid expansion of glacial lakes caused by climate and glacier retreat in the central Himalayas. Hydrol. Process. 2015, 29, 859–874. [CrossRef] 24. Khan, A.J.; Koch, M. Selecting and downscaling a Set of climate models for projecting climatic change for impact assessment in the upper Indus basin (UIB). Climate 2018, 6, 89. [CrossRef] 25. Yu, W.; Yang, Y.-C.; Savitsky, A.; Alford, D.; Brown, C.; Wescoat, J.; Debowicz, D.; Robinson, S. The Indus Basin of Pakistan: The Impacts of Climate Risks on Water and Agriculture; The World Bank: Washington, DC, USA, 2013. 26. Hunt, J.R.; Lilley, J.M.; Trevaskis, B.; Flohr, B.M.; Peake, A.; Fletcher, A.; Zwart, A.B.; Gobbett, D.; Kirkegaard, J.A. Early sowing systems can boost Australian wheat yields despite recent climate change. Nat. Clim. Chang. 2019, 9, 244–247. [CrossRef] 27. Rashid, M.; Haider, S.; Latif, M.; Raja, N. Multi-objective optimization for irrigation deficit through cascade‘reservoirs. Eur. Water 2017, 59, 323–329. 28. Kindler, J.; Tyszewski, S. Multicriteria evaluation of decision rules in the design of a storage reservoir. System analysis for water resources management: Closing the gap between the theory and practice. In Proceedings of the Baltimore Symposium, Baltimore, MD, USA, 19–22 May 1989. 29. Kim, U.; Kaluarachchi, J.J. Climate change impacts on water resources in the upper blue Nile river basin, ethiopia 1. JAWRA J. Am. Water Resour. Assoc. 2009, 45, 1361–1378. [CrossRef] 30. Khan, A.; Richards, K.S.; Parker, G.T.; McRobie, A.; Mukhopadhyay, B. How large is the upper Indus basin? The pitfalls of auto-delineation using DEMs. J. Hydrol. 2014, 509, 442–453. [CrossRef] 31. Mukhopadhyay, B.; Khan, A. A quantitative assessment of the genetic sources of the hydrologic flow regimes in upper Indus basin and its significance in a changing climate. J. Hydrol. 2014, 509, 549–572. [CrossRef] 32. Cook, E.R.; Palmer, J.G.; Ahmed, M.; Woodhouse, C.A.; Fenwick, P.; Zafar, M.U.; Wahab, M.; Khan, N. Five centuries of upper Indus river flow from tree rings. J Hydrol 2013, 486, 365–375. [CrossRef] 33. Tariq, S.M. Pakistan-India Relations: Implementation of Indus-Water Treaty–A Pakistani Narrative; The Pakistan Institute of Legislative Development and Transparency (PILDAT), Office: Islamabad, Pakistan, 2010; p. 9. Sustainability 2020, 12, 7822 18 of 18

34. Ray, P.A.; Brown, C.M. Introduction. In Confronting Climate Uncertainty in Water Resources Planning and Project Design: The Decision Tree Framework; World Bank Publications: Washington, DC, USA, 2015; pp. 1–6. 35. Brown, C.; Steinschneider, S.; Ray, P.; Wi, S.; Basdekas, L.; Yates, D. Decision scaling (DS): Decision support for climate change. In Decision Making under Deep Uncertainty; Springer: Cham, Switzerland, 2019; pp. 255–287. 36. Sieber, J. WEAP water evaluation and planning system. In Proceedings of the 3rd International Congress on Environmental Modelling and Software, Burlington, VT, USA, 1 July 2006. 37. Hassan, D.; Burian, S.J.; Bano, R.; Ahmed, W.; Arfan, M.; Naseer Rais, M.; Rafique, A.; Ansari, K. An assessment of the Pakistan water apportionment accord of 1991. Resources 2019, 8, 120. [CrossRef] 38. Nash, J.E.; Sutcliffe, J.V. River flow forecasting through conceptual models part I—A discussion of principles. J. Hydrol. 1970, 10, 282–290. [CrossRef] 39. Hashimoto, T.; Stedinger, J.R.; Loucks, D.P. Reliability, resiliency, and vulnerability criteria for water-resource system performance evaluation. Water Resour. Res. 1982, 18, 14–20. [CrossRef] 40. Tariq, M.A.U.R.; Van De Giesen, N. Floods and flood management in Pakistan. Phys. Chem. Earth 2012, 48, 11–20. [CrossRef] 41. Kindler, J. The Monte Carlo approach to optimization of the operation rules for a system of storage reservoirs/la méthode Monte Carlo pour l’optimization des règles opératoires pour un système de réservoirs. Hydrol. Sci. J. 1977, 22, 203–214. [CrossRef] 42. Robinson, P.J. Climate change and hydropower generation. Int. J. Climatol. 1997, 17, 983–996. [CrossRef] 43. Harrison, G.; Whittington, H.; Wallace, A. Sensitivity of hydropower performance to climate change. Int. J. Power Energy Syst. 2006, 26, 42. [CrossRef] 44. Zeng, R.J.; Cai, X.M.; Ringler, C.; Zhu, T.J. Hydropower versus irrigation-an analysis of global patterns. Environ. Res. Lett. 2017, 12.[CrossRef] 45. Amin, A.; Iqbal, J.; Asghar, A.; Ribbe, L. Analysis of current and future water demands in the Upper Indus Basin under IPCC climate and socio-economic scenarios using a hydro-economic WEAP model. Water 2018, 10, 537. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).