Te 03 Recursos

8/14/2019 Te 03 Recursos

1/7

UNIVERSIDAD NACIONAL DE INGENIERAFACULTAD DE INGENIERA CIVILDEPARTAMENTO ACADMICO DE HIDRULICA E HIDROLOGA

ANLISIS DE ARTCULOS CIENTFICOS INDEXADOS Pgina1

THIRD STEPPED WORK:INTEGRATED WATERSHED MANAGEMENT MODELING:

GENERIC OPTIMIZATION MODEL APPLIED TO THE IPSWICHRIVER BASIN(ENGLISH VERSION)

Viktoria I. Zoltay, A.M.ASCE; Richard M. Vogel, M.ASCE; Paul H. Kirshen, M.ASCE;andKirk S. Westphal, M.ASCE

RECURSOS HIDRULICOS FACULTAD DE INGENIERA CIVIL

UNIVERSIDAD NACIONAL DE INGENIERA

Abst rac t : A generic integrated watershed management optimization model was developed toefficiently screen a broad range of technical, economic, and policy management options within awatershed system framework and select the optimal combination of management strategies andassociated water allocations for designing a sustainable watershed management plan at least cost.The watershed management model integrates both natural and human elements of a watershedsystem including the management of ground and surface water sources, water treatment anddistribution systems, human demands, wastewater treatment and collection systems, water reusefacilities, nonpotable water distribution infrastructure, aquifer storage and recharge facilities, stormwater, and land use. The model was formulated as a linear program and applied to the upperIpswich River Basin in Massachusetts. Our results demonstrate the merits of integrated watershedmanagement by showing (1) the relative efficacy and economic efficiency of undervalued or

underutilized management options such as incentive pricing; (2) the value of managementstrategies that serve several functions such as the benefits of increased infiltration for meeting bothstorm water and water supply management objectives; and (3) that both human and environmentalwater needs can be met by simultaneously implementing multiple diverse management tools, whichin this case study led to achieving 70% of the recommended in-stream flow with only 25% decreasein net benefits.

DOI: 10.1061/ (ASCE) WR.1943-5452.0000083CE Database subject heading s: Optimization models; Integrated systems; Decision support

systems; Water supply; Watersheds; Water management; Stormwater management; Landmanagement; Wastewater management; Groundwater recharge.

Autho r keywords : Optimization models; Integrated systems; Decision support systems; Watersupply; Watersheds; Water management; Storm water management; Land management;

Wastewater management; Groundwater recharge.

IntroductionIntegrated water resources management

(IWRM) is a rapidly developing fieldencompassing many disciplines includingecology, engineering, economics, and policy.We refer to the Global Water Partnershipsdefinition of IWRM as a process that promotesthe coordinated development and managementof water, land, and related resources in order tomaximize the resultant economic and socialwelfare in an equitable manner without

compromising the sustainability of vitalecosystems (GWP 2009). Historically, waterresources models have been used to facilitatemanagement decisions but they usually dealtwith only a single component of the watershedsystem such as reservoir operations or waterdistribution system design. More recently,IWRM models combine the natural hydrologiccycle with the human water systems technical,socioeconomic, and political components(Jamieson and Fedra 1996; Labadie et al.2000; Zagona et al. 2001; Donigian and Imhoff


2/7


ANLISIS DE ARTCULOS CIENTFICOS INDEXADOS Pgina 2

2002; Fisher et al. 2002; Draper et al. 2003;Letcher et al. 2004; Yates et al. 2005; andothers).

As IWRM models continue to integratevarious aspects of the complex coupled natural-

human watershed system and our engineeringcapabilities continue to develop, we have theability to progress from addressing single

purpose water resources problems to thesimultaneous consideration of joint solutions tomultiple water resources problems by managingthe whole watershed . Consideration of jointsolutions is essential because the integratednature of watershed processes results ininterrelated problems. For example, addressingstorm-water runoff through increased infiltrationnot only reduces peak discharge and improveswater quality in rivers but can also increase theavailability of water supply through the rechargeof aquifers. For planning such integratedmanagement, the appropriate units ofmanagement are watersheds United StatesEnvironmental Protection Agency (US EPA)1996; European Union Water FrameworkDirective (EU WFD) 2000. IWRM applied at thewatershed scale allows for integratedmanagement across disciplines as well asacross the hydrologic cycle and all othercomponents of the watershed system.

For the integrated management ofwatersheds, there is a need for models thatfocus on the development of comprehensivewatershed management models as opposedto the incremental enhancement of existingwatershed hydrologic models. Such models maybe more appropriately referred to as integratedwatershed management models (IWMMs)and we define them as models that (1) aredeveloped for modeling watershedmanagement alternatives with the goal ofunderstanding the effects of managementdecisions on the watershed system in order tosupport decision making and stakeholdernegotiations; ( 2) integrate all relevantcomponents of the natural watershed, humanwater system, and applicable managementtools; and (3) are formulated in a systemscontext, preferably with managementoptimization capabilities to aid in the selectionof promising combinations of managementstrategies. In addition, we advocate thedevelopment of generic models that aretechnically and financially accessible to allowtheir application to watersheds with diversecharacteristics.

In a review of published models, two of themost c omprehensive generic IWRM modelswere water evaluation and planning (WEAP)

(Yates et al. 2005 and WaterWare Jamiesonand Fedra 1996); other reviewed modelsinclude MODSIM (Labadie et al. 2000),RiverWare (Zagona et al. 2001), HSPF(Donigian and Imhoff 2002), and MULINO DSS

(Mysiak et al. 2005; MULINO DSS 2007).However, WaterWare requires specializedexpertise for use and expensive hardware andsoftware support. The cost for the basicWaterWare simulation software is over USD70,000 with over USD 60,000 for the basicoptimization module (WaterWare 2007).Balancing the advantages and hindrances ofcomplex versus simple models is critical notonly for accurate modeling and computationalefficiency but also for usability andtransparency (see Rogers 1978; Ford 2006).Usability includes technical and economicconsiderations and WEAP meets both with asimplified yet accurate model relative toWaterWare and a 2-year licensing costbetween USD 1,000 and USD 2,500(WEAP21 2007). Although WEAP meets thecriteria for generic, comprehensive, integrated,and accessible, it does not providemanagement optimization other than forbalancing water supply reservoir storagecontents.

Since no models were found that met ourobjectives for an IWMM, we introduce a generic

IWMM with optimization capabilities thatefficiently and economically screens a widediversity of options for managing the quantity,quality, routing, timing, and use of waterthroughout a watershed. Sixteen managementoptions are considered simutaneously in orderto account for the positive, negative, direct, andindirect effects of each option in the net benefitcalculation. Such management options areoften modeled independently by the responsiblemanagement agency. Here, we will show thatan IWMM, which simultaneously considersand optimizes all management options within a

watershed framework, can increase theefficiency and effectiveness of resourcesinvested in water resources management.

Model Formulation

Model Schemat ic

The integrated watershed managementoptimization model introduced here is a genericand parsimonious lumped parameter screeningmodel that integrates the natural hydrologiccycle, human water system, and a wide rangeof management options. To accommodate fastsolution times for the future development of an


3/7



interactive decision or negotiation supportsystem, the model is spatially aggregatedtreating the watershed as a single hydrologicresponse unit with a monthly time step over asingle year. We assume that the monthly water

allocations to any system component can berefined to weekly or daily operational valuesthrough detailed simulation modeling. The firstversion of the model is developed for within-year water supply systems that are common inthe Northeastern United States. Thus, weassume that groundwater and surface waterlevels are the same at the beginning and end ofthe year and that a 1-year time horizon isadequate to capture the system dynamics. Themodel was developed in Excel to facilitate thegeneralized application and modification of themodel. The model schematic is shown in Fig.1.

The natural components of the watershedsystem are depicted with white backgrounds.These include land use, runoff, percolation,surface water, groundwater and externalsurface water, and groundwater. Runoff andpercolation are specified as unit values of flowper land area for each land-use type for ahydrologic design condition. These values arecalculated based on precipitation, temperature,and land-use parameters during apreprocessing step using the hydrologicsimulation component of TMDL2K

(Limbrunner 2008) as described below. Theland-use component specifies the existing areaof each land-use type. Surface water,representing rivers and other landscapesources of water, is assumed to have negligiblechannel storage and hence empties completelywithin each time step. Minimum in-stream flowrequirements may be specified on a monthlybasis. Groundwater is the only naturalwatershed component with storage capacity, ofwhich a prespecified fraction determinesbaseflow in the river.

The human components of the watershedsystem are shown with gray and blackbackgrounds in Fig. 1. Gray is used forcomponents that exist and are managed bywater and wastewater utilities in most watersystems. Black is used for components that donot exist or are not actively managed in mostwatersheds. The human system includes areservoir, potable water treatment plant, potabledistribution system, wastewater treatment plant,wastewater collection system, water reusefacility, nonpotable distribution system, septicsystems, aquifer storage and recharge (ASR)facility, potable water users, nonpotable waterusers, interbasin transfer of water and

wastewater, and surface water andgroundwater point sources. The reservoir maybe a single reservoir or the sum of manyreservoirs assumed to be operated together asa single reservoir system. The potable water

treatment plant treats water from surface water,reservoir, or groundwater sources to drinkingwater standards. Potable and nonpotable waterusers may receive inflow from the potablewater treatment plant or through interbasintransfer, which is assumed to supply potablewater. In addition, nonpotable water users mayreceive inflow from the water reuse facilitythrough the nonpotable distribution system.

Wastewater from water users may bedirected to the septic systems, the wastewatertreatment plant, or interbasin transfer based on

user specified input. The wastewater treatmentplant provides secondary wastewater treatmentto meet surface water discharge qualitystandards. Its effluent may be further treated bytertiary wastewater treatment at the water reusefacility. The water reuse facility effluent may bedirected to ( 1) the nonpotable distributionsystem for direct nonpotable reuse; (2) the

ASR facility for recharge and indirect reuse orbaseflow augmentation; and (3) surface waterfor discharge. The ASR facility may receiveinflow from the surface water, reservoir, or waterreuse facility. Pretreatment is required for

surface water and reservoir inflows. There is aone time step (1-month) delay in flows enteringthe groundwater system from the septicsystems and ASR facility, which is a plausibleassumption given the requirements for thedistance from septic systems and ASRfacilities to potable aquifers (US EPA 2004).Surface water and groundwater point sourcesmay be used to model the withdrawal of water,the discharge of wastewater by industrialfacilities, agricultural diversions, or other privateusers.

Relationships between components arebased on the laws of conservation and aremodeled using the continuity equation. Massbalance equations for all components aredescribed in Zoltay (2007).

Incorpo rating Land-Use-Based Managem entOpt ions

Land-use management is an increasinglyrecognized and important element of waterresources management Falkenmark andRockstrm 2006. Land-use management was

integrated into the model to provide the abilityto manage the runoff to recharge ratio andwater quality. The types of vegetative land


4/7



cover and human land use affect both therouting of the water to evapotranspiration,runoff, or percolation and the amount ofpollution in runoff and percolation. Modelingwatershed processes, such as runoff, and

management options, such as storm-waterbest management practices (BMPs), require adaily time step.

The daily hydrologic simulation component ofthe TMDL2K watershed model wasincorporated into the monthly integratedwatershed management optimization model.TMDL2K is a parsimonious lumped watershedmodel, which includes the representation ofBMPs such as detention ponds, bioretention,and swales for storm-water quantity andquality management (Limbrunner 2008). Thespecific BMP that is modeled is determined fromuser specified parameters of contributing area,storage volume, and groundwater and surfacewater outflow constants. Once TMDL2K iscalibrated, its run time in simulation mode is afew seconds.

The IWMM first executes the calibratedTMDL2K model as a preprocessing step toobtain daily runoff and percolation flows perland area per time period for each land-usetype with and without a BMP applied. The dailyunit values are aggregated on a monthlytimescale. Total monthly runoff and percolationvalues are calculated within the IWMMsoptimization routine as the sum of the unitvalues weighted by their area allocations

among existing and BMP applied land uses.This approach allows for the incorporation ofland-use management and storm-water BMPs,

which traditionally require a daily simulationmodel and consequently reduce optimizationefficiency.

Watershed Management Opt ions

The case study guided the choice ofmanagement options included in this initial

model. The management options and theirprimary effects are summarized in Table 1.Background and discussion of thesemanagement options are detailed in Zoltay(2007). In application, the stakeholders inthe watershed will guide the selection ofmanagement options, which is envisioned asan iterative, participatory, and learning processfacilitated by an IWMM such as the oneintroduced here [for participatorymanagement and social learning, see EUWFD ( 2000); US EPA (1996); GWP (2009);Pahl-Wostl (2007)] . The social and politicalsystems within a watershed can significantlyconstrain the acceptability and implementationof some management options. Suchconstraints are reflected in our choices ofmanagement options, for example, byexcluding the direct potable reuse ofwastewater; which is not socially or legallyaccepted in the United States and an upperlimit on the price of water and wastewaterservices.

Land Use and Storm-Water Management

The land-use and storm-watermanagement options enable the optimal


5/7



conversion and routing of precipitation over thewatershed into percolation and runoff, whichaffect downstream water quantity and quality.The land-use management option is (1) landconservation to preserve currently undevelopedforest land at a cost that reflects the purchaseand maintenance of the land. For storm-watermanagement; (2) bioretention units are theBMP considered because groundwaterrecharge is a central concern in the case studyand bioretention BMPs promote infiltration andrecharge. The construction and maintenancecost for bioretention units are based on theirtotal service area.

Minimum and maximum areas for eachland use may be specified for both storm-watermanagement and land conservation options toreflect physical, technical, political, and sociallimits on land-use change. For example, inconsidering land conservation, the area ofexisting urban land limits conversion to anyother type of land use. This can be specified asa minimum urban land area constraint in themodel. A maximum urban area may also bespecified based on existing zoning laws anddevelopment regulations. For storm-watermanagement decisions, a maximum value maybe specified based on the area available forBMP application (i.e., in dense urban settings,physical space is limited for BMPs). Finally,

total land area in the watershed must beconserved through all land area reallocations,and all land area transferred from a regular to a

corresponding BMP land use must not exceedthe original land area for that land use.

Water Supp ly and Treatment

Water supply management options includeincreasing the capacity of the (3) surface waterpumps, (4) groundwater pumps, (5) watertreatment plant and (6) reservoir storage, and(7) the detection and repair of leaks from thedistribution system. The repair of 100% of leaksis not always financially feasible nor does thecost remain linear beyond a certain threshold;hence a maximum feasible repair limit may bespecified. As shown in Fig. 1, leaks from thedistribution system recharge the groundwater;hence, reductions in leaks reduce the recharge.

Human Demand Management

The human demand management optionincludes (8) conservation pricing for water andwastewater services. Conservation pricing isimplemented by specifying a price elasticity foreach water use sector, and the decisionvariable is the percent change in price.

Wastewater Treatm ent

Wastewater treatment management options

include (9) the expansion of secondarywastewater treatment capacity and (10) theconstruction of a water reuse facility with


6/7



tertiary treatment. The water reuse facility maybe a stand-alone facility or an upgrade to theexisting secondary wastewater treatmentplant. Tertiary treated wastewater can bedirected to the surface water and groundwater

for discharge and indirect reuse or to anonpotable water distribution system for directreuse by (11) building a nonpotable distributionsystem. Since the consumption of potable andnonpotable water may be significantly differentdepending on their end use, separateconsumptive use values may be specified. If thenonpotable water use option is implemented, anew percent potable water consumption iscalculated. For example, if the nonpotablewater is used for toilet flushing, anonconsumptive use, then the percentconsumptive use for the remaining uses ofpotable water will change. By including waterreuse as an option, the model is able todetermine when it is economically feasible andefficient to begin tertiary wastewater treatment.Quantifying the benefits of water reuse mayincrease its appeal and acceptance as a watersupply management option. (12) Leak detectionand repair is also available for the wastewatercollection system because groundwaterinfiltration can constitute approximately 40% ofthe wastewater arriving at the treatment plant(MWRA, personal communication, 2007).

Aq uifer Storage and Recharge

(13) The management option to build an ASRfacility allows for the injection of surface wateror treated wastewater into the groundwater.Increased groundwater storage may bebeneficial for restoring groundwater levels,augmenting surface water baseflow orrecovering it for human use.

Interbasin Transfer

Interbasin transfer of (14) water and (15)wastewater can be a desirable managementoption and is included to account for theexisting interbasin transfer of wastewater in thecase study.

Costs and Revenues

All implemented management options incurcosts that may include an initial fixed cost andan annual operations and maintenance cost.Some existing systems such as water treatmentplants have a finite life cycle; thus, theirreplacement cost beyond their life cycle isincluded if the planning period is greater than

the remaining life. The total cost is the sum ofthe total annualized initial cost plus annualoperations and maintenance cost for eachmanagement option. Total revenue is the sumof the user fees from potable and nonpotable

water sales and wastewater services. The netbenefit of watershed management is the totalrevenue minus the total cost. Becausemethodology for quantifying the benefits of in-stream flow and maintaining ecosystemintegrity is not well established, we specified in-stream flow requirements as constraints.

Linear Programm ing Opt im izat ion

The problem was formulated as a linearprogram (LP) to enable efficient use of themodel as a decision support system. Theobjective of the LP optimization is to maximizethe net benefit of watershed management whilemeeting the specified constraints includinghuman demand, management limits on humandemand reduction, in-stream flow standards,land-use restrictions, the capacity or volume offacilities, and surface water and groundwaterflow out of the watershed (see Zoltay (2007)for equations). Frontline Systems premiumsolver platform LP solver was used in Excel tosolve for the values of the decision variables.

Case Study

Backg r o u n d

The upper Ipswich River Basin (IRB) inMassachusetts, which is the watershedupstream of the South Middleton gaugingstation of U.S. Geological Survey (USGS),served as our case study. The Ipswich Riverexperiences low and no flow events at thisgauge during summer months due to heavy

groundwater and other upstream waterwithdrawals. The most recent recommendationsfor seasonal in-stream flow targets for theprotection of aquatic habitat are 0.62 m3/s fromJune to October, 1.25 m 3/s from November toFebruary, 3.14 m3/s from March to April, and1.87 m3/s in May (Zarriello 2002). A detailedhydrologic modeling study of the IRB wasconducted by Zarriello and Ries (2000) of theUSGS using HSPF. That study compiledextensive information on the basin, which wasused in the case study application. Relevantbackground information is summarized below

and the reader is referred to the 2000 study forfurther details.The upper IRB covers approximately


7/7



11,400 ha of land, which is 77% developed and23% undeveloped [Massachusetts GeographicInformation System (MassGIS) 2007]. Itcomprises 14 towns but only four of thesetowns, Reading, North Reading, Wilmington,

and Lynnfield, utilize the upper IRB for theirwater supply. The town of Lynn is not located inthe upper IRB but obtains 16% of its watersupply from it (Zarriello and Ries 2000).Groundwater is the predominant source ofwater in the upper IRB except for the Town ofLynn and supplemental water in the summer forthe Town of Lynnfield. The majority of thewastewater is discharged outside the basin.Reading and Lynn are both on sewer systemsthat export or discharge their wastewateroutside of the basin. Since the other threetowns are on septic systems, it may appear thatthe majority of wastewater is recharged viaseptic systems; however, only North Reading isentirely within the basin boundary. Therefore,even septic systems discharge to other basinsrather than recharging the upper IRB andaugmenting the flow of the Ipswich River.Extensive groundwater withdrawals and theexport of wastewater have been recognized asthe most significant contributors to the low flowevents in the late summer (Zarriello and Ries2000). As the groundwater reserves aredepleted by human withdrawals, the baseflow

of the river is diminished and low and no flowconditions occur from June to August. In somecases, municipal wells are so close to the riverand pump at such a high rate that the riverbecomes the primary recharge source for thewellfield. Zarriello and Ries (2000) estimatedthat (18.9 22.7) x 106 l/day of additionalwater are required in the headwater reaches ofthe IRB to alleviate the low flow events.

The model was applied for 1999, whichwas an average year based on annualprecipitation from 1961 to 2001 but in whichless than a quarter of the in-stream flow target

was achieved during the period of June August. A combination of factors sharplyreduced in-stream flow starting in April including(1) human demand increasing from 0.24 m3/s inJanuary to a summer high of 0.33 m3/s in July;(2) precipitation decreasing from 17.5 cm inJanuary to a summer low of 0.3 cm in June;and (3) increased evapotranspiration. Themodel was applied to the upper IRB to evaluatea broad range of management options formeeting human water demand and therecommended in-stream flow targets.

Model Appl ica t ion

The daily watershed model TMDL2K wascalibrated using 1999 land use, meteorology,streamflow, and groundwater and surface waterwithdrawals for the upper IRB and was used toobtain unit runoff and percolation values (seeZarriello and Ries 2000 for input data). Thegroundwater basin underlying the upper IRBwas assumed to coincide with the surfacewatershed. Human demand was defined as thesum of groundwater and surface waterwithdrawals. Costs associated with theimplementation of management options arelisted in Table 2. The prices for water andwastewater services were $636 per millionliters ($1.80 per 100 ft3) for water and $1,639per million liters ($4.64 per

100 ft3

) for wastewater (MWRA, personalcommunication, 2007). Values for priceelasticity were - 0.2, - 0.2, - 0.5, and - 0.1 forthe residential, commercial, industrial, andagricultural sectors, respectively (Stallworth2003; Olmstead and Stavins 2007). Increase inthe price of services and the resulting decreasein demand did not affect the loss of waterthrough leaks in the distribution system. Costswere compiled from literature as well as directcontact with water and wastewater utilities inthe upper IRB (Zoltay 2007). A 50-yearplanning period and 5% interest rate were usedto compute annual net benefit.Four scenarios were set up formanagement optimization. Results from eachmanagement optimization scenario were compared to the 1999 allocation simulation, ourbase case scenario of what occurred in 1999.The management optimization scenarios wereused to determine management strategies thatcould have been implemented in or prior to1999 to improve watershed conditions. Asummary of which management options wereavailable for each scenario is reflected in Table3 where modeling results are presented.

Te 03 Recursos

Documents

Transcript of Te 03 Recursos