Cooperative Modeling: Linking Science, Communication, and Ground Water Planning

Issue Paper/

Cooperative Modeling: Linking Science,Communication, and Ground Water Planningby Vincent C. Tidwell1 and Cors van den Brink2

AbstractEquitable allocation of ground water resources is a growing challenge due to both the increasing demand for

water and the competing values placed on its use. While scientists can contribute to a technically defensible basisfor water resource planning, this framework must be cast in a broader societal and environmental context. Giventhe complexity and often contentious nature of resource allocation, success requires a process for inclusive andtransparent sharing of ideas complemented by tools to structure, quantify, and visualize the collective understand-ing and data, providing an informed basis of dialogue, exploration, and decision making. Ideally, a process thatpromotes shared learning leading to cooperative and adaptive planning decisions. While variously named, medi-ated modeling, group modeling, cooperative modeling, shared vision planning, or computer-mediated collabora-tive decision making are similar approaches aimed at meeting these objectives. In this paper, we frame‘‘cooperative modeling’’ in the context of ground water planning and illustrate the process with two brief examples.

IntroductionThe solution of environmental problems is no longer

seen as the exclusive responsibility of government; rather,quasi-governmental organizations, the business commu-nity, civic organizations, and citizens are demandinga greater voice in resource planning (e.g., Driessen andGlasbergen 2002). Emergent forms of direct participation(e.g., interactive governance) by interested citizens andother stakeholders in processes of collective decisionmaking are reducing the distance between governmentand their citizenry. The value of collaboration, wherebyvarious stakeholders work with policy makers to addressa particular issue, has been well documented (Connickand Innes 2003; Spash 2001; Claussen 2001; Susskind et

al. 2001). Such collaboration has increasingly includedmodel building to inform the decision process. In fact, theuse of models in open decision-making processes is nota new concept as there are case studies dating back to atleast 1961 (Rouwette et al. 2002). While variouslynamed, mediated modeling (van den Belt 2004), groupmodeling, cooperative modeling (Tidwell et al. 2004),shared vision planning (Palmer 1999), computer-medi-ated collaborative decision making (Kreamer and King1988), the basic idea remains the same—it is a processfor involving stakeholders in the conceptualization, speci-fication, and synthesis of knowledge and experience intouseable information (i.e., model) for the express purposeof addressing a complex problem.

Advantages reported in support of cooperative model-ing include providing participants with system insight,scoping analyses, and education toward a common under-standing of the issues, thus improving the odds that re-sults from the collaborative effort will be implemented indecision making (Palmer et al. 1993; Vennix 1996;Rouwette et al. 2002; van den Brink et al. 2003; van denBelt 2004). Participant values and preferences becomebetter defined, while knowledge levels and consensus onmitigating approaches improve (Costanza and Ruth 1998;Rouwette et al. 2002; van den Belt 2004). Resulting

1Corresponding author: Sandia National Laboratories, P.O.Box 5800, MS 1350, Albuquerque, NM 87185; (505) 844-6025;[email protected]

2Royal Haskoning, Groningen, The Netherlands; [email protected]

Received April 2007, accepted September 2007.No claim to original US government works.Journal compilationª 2008 National Ground Water Association.doi: 10.1111/j.1745-6584.2007.00394.x

174 Vol. 46, No. 2—GROUND WATER—March–April 2008 (pages 174–182)

synergies enable the group to develop new policy optionsin deadlocked processes (van Eeten et al. 2002). Finally,modeling conducted in a cooperative context providesvalue by way of structuring group thinking and dialogue,ultimately leading to group learning (van den Belt 2004).

Criticisms have also been levied against public par-ticipation processes. One argument cites the opportunityfor special interest groups to gain disproportionate influ-ence over the decision process (Coglianese 1999; Kenney2000). Underrepresentation by disenfranchised groupscan lead to inequities in water allocation (World Bank2003). Lord (1986) and Ingram and Schneider (1998)recognize that value disagreements are often masked byfactual disputes in public debate. The factual disputesthen become the focus of the process and the situation ofadversarial science or ‘‘dueling experts’’ arises. Coopera-tive modeling efforts may also encourage unrealistic ex-pectations; that is, no level of modeling can make up forlack of understanding of basic principles—an inadequatemodel will never provide a sound basis for decision mak-ing or course of action. Finally, collaborative processesare blamed for taking longer and costing more to reacha decision, although this point is questionable where theissue results in litigation.

Prior to engaging in any cooperative modeling pro-cess, careful attention must be paid to the decision land-scape. Toward this end, our objective is to identify keyconsiderations in the design and implementation of a coop-erative modeling project, specifically framed in the contextof ground water planning. Given the breadth of this subject,our intent is simply to raise awareness, while steering theinterested reader to the literature for details on ‘‘how to.’’

We begin by discussing the problem domain ofground water planning, particularly the unique character-istics and challenges it poses for cooperative modeling.We then define and discuss three key considerations inthe design of a cooperative modeling project: public par-ticipation, modeling, and their fusion. Finally, we pro-vide a few brief case studies that highlight these designpoints.

Ground Water PlanningAlthough not entirely unique, there are several fea-

tures inherent to ground water systems that present realchallenges for resource planning. Uncertainty in thequantitative characteristics of the ground water systemand potable resource in storage is a leading source ofcontention. Uncertainty stems from the general difficultyin measuring broad distributed characteristics of thesystem such as basin bathymetry, aerial recharge, river/aquifer interaction, and evapotranspiration losses. Addi-tionally, our view of the ground water system is limited;that is, boreholes, which sample only a small fraction ofthe entire aquifer, are often our only portals to measuringkey petrophysical and water quality characteristics. Fur-ther, complicating matters is the spatial and temporal het-erogeneity inherent to the physical characteristics of theground water system as well as in land-use function andmanagement. While aquifer characterization technology

is rapidly improving, costs to adequately resolve keymeasures within the context of resource planning are gen-erally prohibitive, particularly at levels capable of illus-trating the role of individual stakeholders within theproblem and the interrelation between parties. Ultimately,the danger is uncertainty leading to dueling models andpolarization of views by competing interests.

Another challenge relates to the ‘‘mysterious’’ natureof ground water systems. The poorly informed may comewith visions of underground rivers and lakes, as well asideas of ‘‘infinite ground water supplies.’’ While thebetter informed may hold more physically based mentalmodels, their conceptualizations are often biased towarda particular interest. Thus, the challenge is to help all par-ticipants to learn toward a shared view of the groundwater system that in turn is scientifically defensible.

Not only are ground water systems complex in theirown right but also the dynamics of these systems areinextricably linked to many other physical and social sys-tems. Resource planning requires an analysis adjusted notonly to the nature of the ground water issues faced butalso considering social, economic, and legal constraints(Alley and Leake 2004). In this broader context, decisionstranscend direct technical arguments to include valuejudgments. Thus, equitable resource planning defies myo-pic, piecemeal approaches but rather requires the fusionof knowledge and experience distributed across disciplin-ary and constituency domains.

Cooperative ModelingBefore we tackle structuring of a cooperative model-

ing project, we first need to consider when such anapproach is appropriate. Although not a comprehensivelist, we offer four considerations toward this decision.First, there must be willingness on the part of key interestgroups to cooperate in the planning process. In situationsinvolving entrenched, high-intensity conflict, the stakesare often described in narrow terms, thus leaving littleroom for dialogue and compromise (van den Belt 2004).Second, each interest group needs equal access to theplanning process; that is, participation cannot be un-equally constrained by proximity, financial resources, orthe like. Third, the entity responsible for conducting thecooperative modeling process must be trusted by all inter-ests. The entity must be viewed as an unbiased facili-tator working toward the collective benefit of the region.Finally, the availability of financial resources must be fac-tored into the decision.

Public ParticipationIn a perfect world, all interested parties would be

invited to participate in a cooperative modeling process;however, this is rarely practical. The challenge arisesbecause the cooperative modeling process strives forbroad participation, while meeting efficiency and produc-tivity is best achieved in groups limited to 10 to 30 partic-ipants (van den Belt 2004). Fortunately, participation canbe balanced with efficiency through careful structuring ofthe public participation process.

V.C. Tidwell, C. van den Brink GROUND WATER 46, no. 2: 174–182 175

‘‘Circles of influence’’ (Institute for Water Resources1994) is one such approach that organizes participationbased on level of interest and ability to engage. Theapproach can be envisioned as a series of concentric cir-cles with the innermost circle encompassing the ‘‘supportteam.’’ By support team, we mean the personnel responsi-ble for physically constructing the model, process facilita-tion, and project management. The next larger circleincludes the most interested stakeholders who througha combination of time, energy, and aptitude are involvedin the technical work. This ‘‘cooperative modeling team’’(CMT) directly engages in the development and/or appli-cation of the model and thus develops a detailed knowl-edge of the technical process. Ideally, each participantacts as the liaison for a broader interest group. A thirdouter circle is drawn to encompass all other stakeholderswho desire only a peripheral role in the process. Thefourth and final circle contains the political decisionmakers that provide direction to the effort. This circleoverlays the three concentric circles, with informationexchange and dialogue occurring directly with each.

Here, an interest group or individual stakeholder isanyone with a stake in the decision. In this context, an in-terest group might include government organizations, publicservice organizations, nongovernment organizations, orbusiness organizations, while an individual stakeholdermight include representatives of a private business, tech-nical subject experts, or a concerned citizen. Involvementin the outer third circle is simply a self-selection processas such participation is open to all. Participation in theCMT generally occurs by a hybrid self-selection andappointment process. Ideally, each interest group wouldelect a representative, who in turn is willing and able toparticipate. The guiding principle in participant selec-tion is to keep the process as open as possible whileworking to keep the representation of the individual inter-ests balanced.

Beyond the cast of stakeholders and interest groupsis the support team. This is the group that prepares themeetings, guides the discussion, synthesizes group think-ing, and simultaneously translates the results into a mod-el. This involves the four distinct roles of facilitator,modeler, process coach, and gatekeeper (Richardson andAndersen 1995). The gatekeeper is the project championwho is the promoter and energy behind the project.Although an individual can fulfill more than one role, it isimportant to recognize the need for each and deliberatelyassign responsibility to skilled participants.

It is important to establish ground rules at the begin-ning of the cooperative modeling process to providea structure for group dynamics (Wondolleck and Yaffee2000). Participants must agree on rights and responsibili-ties, particularly attendance and participation expect-ations. Behavior guidelines should be set. Clear rulesmust also be set on how group decisions will be han-dled—by consensus, majority vote, or some hybrid. Suchguidelines are necessary before contentious decisions areencountered. It is also important to define expectationsfor the modeling exercise, specifically the purpose androle the model will play in any future decision process(e.g., Tidwell et al. 2004).

ModelingCollaboration is pursued with the goal of develop-

ing broad consensus within a decision context; how-ever, consensus in the absence of scientifically soundunderstanding is of limited value. Modeling plays a cen-tral role in establishing a sound and informed basis fordecision making and provides a vehicle for communicat-ing technical results.

In cooperative modeling, it is modeling that connectsthe decision process with science. Specifically, modelingforces participants to confront facts. ‘‘Science has de-veloped specific methods and peer review processes tomaintain as objective a view as possible, but in other envi-ronments this conscious effort is often lacking’’ (van denBelt 2004). The process of jointly developing and exercis-ing a model also helps the group to better appreciate sys-tem complexities and understand its cause and effectrelations (Vennix et al. 1997). Additionally, models pro-vide a uniform basis for comparing competing alternativesand assessing tradeoffs. The aim of cooperative modelingis not to provide a scientific evaluation of a ground waterissue but to identify the best course of action to resolvestakeholder conflict (van den Brink et al. 2007).

Models can be used in a variety of modes within thedecision process. At one end of the spectrum, models areused only in an illustrative capacity to help participantsvisualize key aspects of the system. At the opposite end,models are used to quantitatively distinguish between com-peting management scenarios. Other roles might includesystem insight, scoping analyses, education, and outreach.

Selecting an appropriate model for a cooperativemodeling project is really no different than that encoun-tered with other modeling exercises. Careful attentionmust be paid to matching model fidelity and resolutionwith model purpose, available data, and modeling resour-ces (e.g., Bear et al. 1992; Anderson and Woessner 1992).What is different is consideration of the means by whichproject participants interact with the model. Participantsneed an environment in which they can query the modeland receive rapid feedback to promote dialogue andshared understanding. Participant interaction can be indi-rect where experts are responsible for manipulating andrunning planning scenarios requested by the team. Alter-natively, the modeling framework can be designed fordirect interaction where participants are able to set up andrun planning scenarios themselves.

To achieve this level of interactivity generally re-quires use of more than just a model, that is, more thana mathematical representation of a real-word system.Rather, the model needs to be accompanied by tools tohelp stakeholders access information and improve under-standing. In particular, a model interface is required,which is the portal by which the expert or participantstructures model input to simulate future desired con-ditions. An interface is also needed to assist with thepresentation and visualization of model results. Otheruseful tools include database management functions anddecision analysis support. Model interactivity is alsoenhanced where model results can be accessed rapidly.This can be accomplished in a few ways. First, modelswith short (tens of seconds) computation times can be

176 V.C. Tidwell, C. van den Brink GROUND WATER 46, no. 2: 174–182

applied. Second, planning scenarios can be simulated priorto a public meeting and stored in a database coupled witha user interface.

Decision support systems (DSS) provide a valuablemodeling framework with which to enhance model inter-activity. While defined in many different ways, a DSScan be taken as ‘‘a computer-based system that aids theprocess of decision making’’ (Finlay 1994). Basic DSScomponents often include a user interface, database, sim-ulation and analysis tools, and the overarching systemnetwork architecture (Power 1997). The required level ofDSS complexity simply depends on the functionalityrequired to connect stakeholder and science in the contextof the particular planning issue.

Negotiation support systems (NSS) and planning sup-port systems (PSS) are types of DSS with application toa particular problem. The spectrum of NSSs include(Starke and Rangaswamy 1999) (1) expert systems that useaccumulated knowledge to aid stakeholders preparing fornegotiations; (2) a system that combines technologies forindividual DSS and group decision support and facili-tates the actual negotiation process in multiagent settings(Nandalal and Simonovic 2003); and (3) ‘‘autonomousagents’’ that are programmed to negotiate on behalf of theirhuman principals. The real advantage of NSSs is that theyaid negotiators in overcoming their cognitive limitationsand in identifying their (and others’) real interests ratherthan focusing on negotiating positions (see also Fisher andUry 1983). A PSS generates information that can be com-municated to and among the stakeholders and used in theplanning process (Geertman and Stillwell 2003).

System dynamics provides yet another platform forcooperative modeling (van den Belt 2004; Winz andBrierley 2007). The process by which system dynamicsmodels are conceptualized and quantified lends itself wellto cooperative modeling (e.g., Sterman 2000). The valueof system dynamics is in capturing the feedback and timedelay between multiple interacting systems, which is keyto effective resource planning. Additionally, availablecommercial software packages provide an integrated soft-ware environment, including object-oriented modeling,database connection, user interface development, anddecision analysis tools.

FusionPreviously, we considered public participation and

resource modeling within the context of cooperative mod-eling. Now, we turn our attention to the fusion of thesekey elements; specifically, we consider when and how tomerge public participation with modeling.

Transparency and trust are key objectives of coop-erative modeling. Building a model from the ground upimproves transparency and trust in the resulting tool.Engaging stakeholders in the actual development of themodel helps improve their understanding of the system,data sources, key assumptions, and limitations. Alterna-tively, the modelers benefit from the experiential knowl-edge of the stakeholders. The broader perspectiveafforded by cooperation also benefits the overall model-ing and decision process. However, it is not always neces-sary for stakeholders to be engaged in the model

development stage. Where tools exist that are widelytrusted, stakeholder involvement can be limited to modelapplication and analysis exercises.

Stakeholder participation primarily occurs throughattendance of group meetings. As a very rough rule ofthumb, a cooperative modeling project will take 60 to80 hours of group meeting time (van den Belt 2004);however, this figure can vary significantly depending onthe complexity of the issue. These meetings are used tocollaboratively define the purpose, scale, and scope of themodel; establish decision variables and metrics; conceptu-alize and structure system cause and effect relations; quan-tify key processes; and specify model data and parametervalues. ‘‘Coding’’ of the model can be performed in themeetings (as is often done with system dynamics model-ing) or performed between meetings by members of thesupport team. Meeting time should also be scheduled forcalibration, testing, and application of the model. Fre-quency and duration of the meetings will depend largelyon the level of participant engagement and their schedules.Cooperative modeling exercises can be pursuedin a compact manner where stakeholders devote severalconsecutive days, or meetings can be dispersed in timemeeting monthly or bimonthly for several hours at a time.

Finally, in structuring a cooperative modeling pro-cess one should carefully consider the nature of theproblem at hand and the underlying source of tension.There are three general groupings, the choice of whichwill influence the fusion of the participant, modeling, anddecision process:

d Complexity of the policy field: policy fields are character-

ized by a complex web of actors serving their interest in

a ‘‘multiactor governance,’’ ‘‘multilevel governance,’’ and

‘‘multisector governance’’ environment. In such a context,

cooperative modeling is helpful in surfacing relevant and

adequate information and putting it into the hands of stake-

holders and decision makers (Sehlke and Jacobson 2005).d Complexity of the physical system: physical systems that

have cause and effect separated by time and or space, that

is, dynamic complexity (Senge 1994). For such systems,

cooperative modeling can improve awareness and under-

standing of the problem, assist in rephrasing of the problem

such that it illustrates the unique role of each stakeholder

within the problem, and illustrate the interrelation between

concerned parties and their shared fate in the dilemma.d Uncertainty in the process: the course and outcome of

a (negotiation) process should leave enough space to iden-

tify the best course of action rather than to communicate

the solution of a major actor.

It is also often advantageous to structure the coopera-tive modeling process toward investigating future trends inkey system behavior rather than attempting to predictthe exact future state or scientific evaluation of a system(Nandalal and Simonovic 2003; van den Brink et al. 2007).

Case StudiesThere are several examples of cooperative modeling

in the context of environmental resource planning,


including assessing the effects of sheep grazing on sagegrouse populations (van den Belt 2004); energy use iniron and steel production (Costanza and Ruth 1998);air quality issues (Stave 2002); sustainability of arcticcommunities (Nicolson et al. 2002); park management(Videira et al. 2003); water planning (Moxey and White1998; van Eeten et al. 2002); and ground water manage-ment (McPhee and Yeh 2004). Here, we offer two addi-tional case studies that address issues important to groundwater planning and that illustrate several of the coopera-tive modeling design aspects noted previously. The firstexample addresses coupled land use–ground water qualitymanagement for a community in the Netherlands, the sec-ond example considers sustainable water planning for athree-county region in north-central New Mexico. Whilethese examples are similar in their use of cooperativemodeling, they differ in scope, geography, modeling plat-forms, and the collaborative framework.

Ground Water Quality ManagementThe Holten region is located in the eastern part of

the Netherlands in the province of Overijssel. This regionenjoys a rural setting with a balance of small municipali-ties, farms/dairies, and open public lands. The local watercompany in cooperation with the provincial authority(responsible for ground water quality and with a clearmandate to resolve the conflict) initiated a local planningprocess in response to mounting concerns over nitrateconcentrations in abstracted and shallow ground water.Drivers for this local process were requirements of theEuropean Commission (EC) Nitrate Directive (EC 1991).In addition to this local planning process, the province isformulating a regional plan in response to the Dutch leg-islation for restoration of areas impacted by intensivelivestock farming (Reconstruction Law). The purpose ofthis legislation is to reach a new balance between varioussocial, economic, and ecologic functions in rural areas.To reach a final agreement, assistance of procedures ormethodologies acceptable to all concerned parties is verymuch needed (Nandalal and Simonovic 2003).

In support of these planning processes, Royal Has-koning and Geological Survey of the Netherlands (TNO)(i.e., the support team) were requested to develop an NSSto assist in solving disagreements among stakeholders.The NSS was used to model and evaluate the interrela-tionships of different components and activities withinthe ground water–land-use system and can be consideredan object-oriented modeling approach (Nandalal and Si-monovic 2003). Purposes of the NSS were to identify thebest course of action to resolve the stakeholder-acceptedconflict between land-use management and ground waterquality; rephrase the problem such that it illustrates theunique role of each stakeholder and the interrelationbetween concerned parties; and explain the complexity inthe ground water system in terms of recharge area andresidence times.

In order to build acceptance and confidence in theNSS, parties including water boards, provincial authori-ties, and water companies (i.e., the CMT) participated ina series of 2-day-long workshops to define the require-ments and select the modeling tools of the NSS. Upon its

completion, a spatial bottom-up planning process wasorganized at the local level. This involved the water com-pany, provincial government, and local stakeholders (i.e.,all other and possibly conflicting stakeholders). Severalmeetings were held over a 9-month period of time toexplore alternative land-use scenarios aimed at reducingnitrate loading to the ground water system. A broader setof agricultural representatives, recreational entrepreneurs,and civil servants of the municipalities became part ofa regional planning process.

Several land-use scenarios were considered, of whichthe ‘‘base’’ and the ‘‘adapted’’ land use were developed toidentify the best course of action to resolve the stake-holder-accepted conflict. The base scenario calculates theautonomous development assuming no changes in theland-use pattern. The adapted land-use scenario is iden-tical to the base scenario until 2005 at which time theadapted land-use scenario was assumed to be imple-mented. The adapted land-use scenario was defined bythe water company and the provincial government basedon (1) the calculated quality of the shallow ground waterin 2005 obtained from the base scenario and (2) the resi-dence times of the ground water toward the abstractionsite calculated with the calibrated geohydrological flowmodel. Experiential knowledge of the stakeholders wasused to set parameter values reflecting historical, current,and future nutrient application for the three agriculturalland uses (grass, maize, and arable land).

The focus of the local planning process was theprotection of drinking water through modification ofland-use practices. The NSS supported this process byproviding a vehicle with which to explore alternative land-use scenarios and providing quantitative information onthe impact of the proposed land-use change and thechange in N application on the quality of abstractedground water. The output generated by the NSS was orga-nized in the form of maps and time series as requested bythe stakeholders. This information visualized and quanti-fied the conflicting interests between agricultural landuse and drinking water function, indicating that sustain-able protection of the resource would require a substantialchange in land use. The changes in nitrate concentrationdue to the adapted land-use scenario are significant asthey are realized in an area with rather short residencetimes (Figure 1). Maps of shallow ground water qualitynot only contributed to the quantification of the land-useconflict but also helped to overcome the difference inplanning horizon of the stakeholders and the time scale ofthe geohydrological system. Based on this insight, thestakeholders realized their own role as well as their com-bined role with other stakeholders in these conflicts. Asa result, the need for a substantial change in the actualland-use pattern to protect the abstraction site becamea shared notion among the stakeholders. In this way,information from the NSS was used in the bottom-up spa-tial planning process as the technical basis for manage-ment decision making (cf. Simonovic 2002).

While formalization of the adapted land-use scenariowas the end point of the local planning process, issuesstill remained. Proposed land-use changes were largelyfocused on agricultural practices, but the local planning


process did not have the means to compensate the farm-ers. To resolve this issue, attention was directed at theregional planning process, which provided a compensa-tion mechanism for farmers located outside the drinkingwater abstraction recharge area. Although the water com-pany and the province were leading actors in the localplanning process, they were just a part of a larger group(including socioeconomic actors) in the regional recon-struction process. This process included political decisionmakers. By building on experience gained in the local

planning process, the water company and the provincewere able to have a strong voice in regional planning.Specifically, insight into the relation between land use,ground water residence times, and ground water qualityas provided by the NSS helped push the local planningprocess scenarios to the forefront of the regional planningprocess. Cooperative modeling proved helpful for inte-grating complex hydrologic data with other information(e.g., policy, regulatory, and management criteria) indefining a course of action to resolve stakeholder conflict

Figure 1. Land use, residence times, and calculated nitrate concentrations (mg NO3/L) in shallow ground water in the basescenario (2005) and the changes in land use and calculated nitrate concentration in the adapted land-use scenario.


(cf. Sehlke and Jacobson 2005). A complete descriptionof this case study can be found in van den Brink et al.(2007).

Regional Water PlanningThe New Mexico Office of the State Engineer

(NMOSE) initiated a statewide water planning process inthe mid-1990s in response to mounting concerns over waterissues in the state. Sixteen planning regions were formed,each with the responsibility of preparing a 50-year watermanagement plan that balances sustainable water supplywith projected demand in a publically and politically ac-ceptable manner. One of the 16 regions, the Middle RioGrande (MRG), adopted a cooperative modeling process tosupport their water planning efforts (Tidwell et al. 2004).This region encompassed three counties in arid north-centralNew Mexico that are centered on the Rio Grande and in-cludes Albuquerque, the principal urban center in the state.

Public participation within the cooperative modelingprocess was largely structured around the circles of influ-ence approach discussed previously. In the inner circlewas the support team, which consisted of two modelersfrom Sandia National Laboratories and a professionalfacilitator provided by the Utton Transboundary Resour-ces Center of the University of New Mexico. In the nextcircle were members of the CMT. Team members wereappointed by the Mid Region Council of Governments(MRCOG), the local government association with over-sight responsibility for the planning process, and the Mid-dle Rio Grande Water Assembly (MRGWA), thevolunteer organization that spearheaded the planning.Team members were selected according to their willing-ness to participate and the need to balance the interests oflocal government, irrigators, environmentalists, urban de-velopers, and regional water managers. The third circleof participants was made up of other MRGWA andMRCOG members not on the CMT, the interested public,and local/state/federal water agency representatives.

The support team in close collaboration with theCMT worked to develop a decision support model to aidthe water planning process. The model was developedfrom the ground up and took approximately 18 months tocomplete. During this period of time, the support andCMTs met on a biweekly basis for 2 to 3 hours per meet-ing. During these meetings, team members helped con-ceptualize the physical system, identify data sources,quantify causal relations, design the interface, and reviewthe model. Coding and calibration of the model waslargely performed outside the meetings by the supportteam modelers. In this way, the CMT was in full controlof the design and implementation of the model and hadfull knowledge of the limitations, assumptions, and un-certainties inherent to the model and associated data. Theouter circle of participants also assisted with modeldevelopment. On a quarterly basis, public meeting wereheld to discuss progress on the model and to solicit feed-back. Additionally, numerous meetings were held withlocal, state, and federal water agency personnel to benefitfrom their expertise, data, and review of the process.

The decision support model was tasked to quantita-tively evaluate alternative water conservation strategies,

engage the public in the planning process, and explaincomplexities of the regional water system. Additionally,the model was designed to be fully accessible to the pub-lic, that is, operate on a PC without the aid of an expertmodeler. Toward these needs, a decision support modelwas created within a system dynamics framework usingthe object-oriented modeling package Powersim Studio2003. Basic model elements include surface water andground water supplies balanced against municipal, agri-cultural, evaporative, and riparian demand. Also includedin the model are 24 water conservation strategies that thepublic identified through community meetings held aspart of the planning effort. The model is accessiblethrough a user-friendly interface that spans approximately80 computerized pages, including pictures, explanatorytext, 66 slider bars, and buttons for exploring alternativewater conservation strategies, and graphs and tables forvisualizing simulation results. Simulations take less than10 seconds to complete.

In spring 2003, a working model was delivered tothe MRGWA and used to develop scenarios that integratevarious combinations and intensities of the 24 conser-vation alternatives. Ultimately, a series of five scenariosintended to represent perspectives from each of fiveconstituency groups (agriculture, environment, urban,specialists, managers) were developed. These scenarioswere then vetted with the public to gather their preferen-ces and perspectives. Working closely with the MRCOG,the MRGWA used the decision support model to combinethe individual scenarios along with public feedback intoa preferred scenario, which then became the basis for theregional water management plan. Once the public hadcommented, the plan was finalized and submitted. TheNMOSE formally accepted the plan in August 2004. Cur-rently, some articles of this plan are in the process ofbeing implemented, while others are being studied anddiscussed in more detail.

Postproject interviews were performed with the gen-eral public and CMT members to gauge perceptions onmodeling, identify strengths and weaknesses of the coop-erative modeling process, and determine if and how themodel facilitated the planning process. Results over-whelmingly support using models in developing solutionsto complex public policy issues and using models to edu-cate and engage the public in such processes. CMT teammembers believed that it was appropriate to use themodel in planning and that the planning process wouldhave been more difficult without it. This is not to say theprocess did not face its share of challenges. Specific les-sons learned include the need for expert facilitation in thecollaborative process, transparent and clear communi-cation throughout the process, and careful definition ofthe role of modeling in the planning process. Additionaldetails can be found in Cockerill et al. (2004, 2006).

ConclusionsWater touches our lives in many ways, from basic

physical needs to the food we eat, the economy, and theenvironment in which we live. Management of this


tightly coupled system benefits from the sharing ofknowledge and experience across disparate disciplinesand constituencies. To assist in the sharing of informationand ideas, models are needed to structure group dialogueand thinking toward shared learning. Models also helpground the process with a scientifically informed basisfor decision making. In this paper, we present a generalframework for cooperative modeling, specifically, a pro-cess for involving stakeholders in the conceptualization,specification, and synthesis of knowledge and experienceinto useable information (i.e., model) for the express pur-pose of addressing a complex problem. This approach isnot predicated on a single decision process or modelingplatform; rather, it encompasses a continuum of opera-tional modes that can be tailored to the unique needs ofan individual problem.

To demonstrate this approach, two examples in thecontext of ground water planning are given, which aresimilar in their use of cooperative modeling, yet differ inscope, geography, modeling platforms, and the collabora-tive process. These examples show that beyond a model’stechnical contribution, there is the added value ofimproved awareness and understanding of the problem,a rephrasing of the problem such that it illustrates theunique role of each stakeholder within the problem andillustrates the interrelation between concerned parties andtheir shared fate in the dilemma. Cooperative modelingtherefore provides a vehicle for identifying the bestcourse of action to resolve ground water conflict.

While cooperative modeling holds promise, it shouldnot be assumed to be the solution to all environmentalconflict. Where there is no willingness to cooperate, liti-gation may be the only alternative. Even where coopera-tive modeling approaches are employed, some level ofcontinued conflict should be expected. In fact, some con-flict is healthy as it suggests that the hard issues are beingaddressed and that competing views are being expressedand discussed. While cooperative modeling does not pre-clude conflict, it does provide a structured process fortransparently and scientifically dealing with the conflict.

AcknowledgmentsThe authors would like to acknowledge the support

and technical contribution of Kristan Cockerill, HowardPassell, Suzanne Pierce, Willem Jan Zaadnoordijk, JasperGriffioen, and Peter de Ruiter toward the development ofthis document. We also sincerely thank participants fromthe Holten and MRG planning processes, for withouttheir contribution there would be no cooperative model-ing. We also acknowledge the very helpful and construc-tive comments received from three anonymous reviewers.Sandia is a multiprogram laboratory operated by SandiaCorporation, a Lockheed Martin Company, for the U.S.Department of Energy’s National Nuclear SecurityAdministration under contract DE-AC04-94AL85000.

ReferencesAlley, W.M., and S.A. Leake. 2004. The journey from safe yield

to sustainability. Ground Water 42, no. 1: 12–16.

Anderson, M.P., and W.W. Woessner. 1992. Applied Ground-water Modeling Simulation of Flow and Advective Transport.San Diego, California: Academic Press.

Bear, J., M.S. Beljin, and R.R. Ross. 1992. Fundamentals ofground-water modeling. EPA/540/S-92/005. Washington, D.C.:U.S. Environmental Protection Agency Office of Researchand Development.

Claussen, E. 2001. Making collaboration a matter of course: Anew approach to environmental policy making. Environ-mental Practice 3, no. 4: 202–205.

Cockerill, K.C., H. Passell, and V.C. Tidwell. 2006. Cooperativemodeling: Building bridges between science and the pub-lic. Journal of American Water Resources Association 42,no. 2: 457–471.

Cockerill, K., V.C. Tidwell, and H. Passell. 2004. Assessingpublic perceptions of computer-based models. Environ-mental Management 34, no. 5: 609–619.

Coglianese, C. 1999. The limits of consensus. Environment 41,no. 3: 28–33.

Connick, S., and J.E. Innes. 2003. Outcomes of collaborativewater policy making: Applying complexity thinking toevaluation. Journal of Environmental Planning and Man-agement 46, no. 2: 177–197.

Costanza, R., and M. Ruth. 1998. Using dynamic modeling toscope environmental problems and build consensus. Envi-ronmental Management 22, no. 2: 183–195.

Driessen, P.P.J., and P. Glasbergen. 2002. New directions inenvironmental politics—Concluding remarks. In GreeningSociety. The Paradigm Shift in Dutch EnvironmentalPolitics, ed. P.P.J. Driessen and P. Glasbergen, 245–262.The Hague: Kluwer Academic Publishers.

EC. 1991. Council Directive 91/676/EEC of 12 December 1991concerning the protection of waters against pollutioncaused by nitrates from agriculture resources. NitrateDirective. OJL 375, p. 0001-0008. Brussels, Belgium: Offi-cial Journal of the European Communities.

Finlay, P.N. 1994. Introducing Decision Support Systems. Cam-bridge, Massachusetts: Blackwell Publishers.

Fisher, R., and W. Ury. 1983. Getting to Yes: Negotiating Agree-ments without Giving In. New York: Penguin Books.

Geertman, S., and J. Stillwell. 2003. Planning Support Systemsin Practice, 1st ed. Berlin, Germany: Springer Verlag.

Ingram, H., and A. Schneider. 1998. Science, democracy, andwater policy.Water Resources Update no. 113: Autumn.

Institute for Water Resources. 1994. Managing water fordrought. IWR Report 94-NDS-8. Alexandria, Virginia: U.S.Army Corps of Engineers.

Kenney, D.S. 2000. Arguing about Consensus: Examining theCase Against Western Watershed Initiatives and Other Col-laborative Groups Active in Natural Resources Manage-ment. Boulder, Colorado: Natural Resources Law Center.

Kreamer, K.L., and J.L. King. 1988. Computer-based systemsfor cooperative work and group decision making. ACMComputing Surveys 12, no. 2: 115–146.

Lord, W. 1986. Social and environmental objectives in waterresources planning and management. In An EvolutionaryPerspective on Social Values, ed. W. Viessman andK. Schilling, 1–11. New York: American Society of CivilEngineers.

McPhee, J., and W.W.-G. Yeh. 2004. Multiobjective optimiza-tion for sustainable groundwater management in semiaridregions. Journal of Water Resources Planning and Man-agement 130, no. 6: 490–497.

Moxey, A., and B. White. 1998. NELUP: Some reflections onundertaking and reporting interdisciplinary river catchmentmodeling. Journal of Environmental Planning and Man-agement 41, no. 3: 397–402.

Nandalal, K.D.W., and S.P. Simonovic. 2003. Resolvingconflicts in water sharing: A systemic approach. WaterResources Research 39, no. 12: 1362–1373.

Nicolson, C.R., A.M. Starfield, G.P. Kofinas, and J.A. Kruse.2002. Ten heuristics for interdisciplinary modeling proj-ects. Ecosystems 5, no. 4: 376–384.


Palmer, R. 1999. Modeling water resources opportunities, chal-lenges, and tradeoffs: The use of shared vision modelingfor negotiation and conflict resolution. In Proceedings ofthe ASCE’s 26th Annual Conference on Water ResourcesPlanning and Management, Tempe, Arizona. Reston,Virginia: American Society of Civil Engineers.

Palmer, R.N., A.M. Keyes, and S. Fisher. 1993. Water manage-ment in the 90’s: A time for innovation. In Proceedingsof the 20th Anniversary Conference, ed. K. Hon, 451–454.Reston, Virginia: American Society for Civil Engineering.

Power, D.J. 1997. What is a DSS? The On-Line Executive Jour-nal for Data-Intensive Decision Support 1, no. 3.

Richardson, G.P., and D.F. Andersen. 1995. Teamwork in group mod-eling building. System Dynamics Review 11, no. 2: 113–137.

Rouwette, E.A., J.A.M. Vennix, and T. van Mullekom. 2002.Group model building effectiveness: A review of assess-ment studies. System Dynamics Review 18, no. 1: 5–45.

Sehlke, G., and J. Jacobson. 2005. System dynamics modelingof transboundary systems: The bear river basin model.Ground Water 43, no. 5, 722–730.

Senge, P.M. 1994. The Fifth Discipline: The Art & Practice ofthe Learning Organization. New York: Bantam DoubledayDell Publishing Group Inc.

Simonovic, S.P. 2002. Two new non-structural measures forsustainable management of floods. Water International 27,no. 1: 38–46.

Spash, C.L. 2001. Broadening democracy in environmental pol-icy processes. Environment and Planning C: Governmentand Policy 19, no. 4: 475–481.

Starke, K., and A. Rangaswamy. 1999. Computer-MediatedNegotiations: Review and Research Opportunities, Encyclo-pedia of Microcomputers, vol. 26. New York: Marcel Inc.

Stave, K.A. 2002. Using system dynamics to improve publicparticipation in environmental decisions. System DynamicsReview 18, no. 2: 139–167.

Sterman, J.D. 2000. Business Dynamics, Systems Thinking andModeling for a Complex World. Boston, Massachusetts:McGraw-Hill.

Susskind, L.E., R.K. Jain, and A.O. Martyniuk. 2001. BetterEnvironmental Policy Studies: How to Design and ConductMore Effective Analyses. Washington, D.C.: Island Press.

Tidwell, V.C., H.D. Passell, S.H. Conrad, and R.P. Thomas.2004. System dynamics modeling for community-based

water planning: An application to the Middle Rio Grande.Journal of Aquatic Sciences 66, no. 4: 357–372.

van den Belt, M. 2004. Mediated Modeling: A System DynamicsApproach to Environmental Consensus Building. Wash-ington, D.C.: Island Press.

van den Brink, C., W.J. Zaadnoordijk, B. van der Grift, P.C. deRuiter, and J. Griffioen. 2007. Using a groundwater qualitynegotiation support system to change land-use managementnear a drinking-water abstraction in the Netherlands. Jour-nal of Hydrology. Accepted.

van den Brink, C., W.J. Zaadnoordijk, J. Griffioen, and B. vander Grift. 2003. NSS for harmonizing land use functionsand sustainable drinking water production. In FramingLand Use Dynamics: Reviewed Abstracts, InternationalConference, 16–18 April 2003, ed. M. Dijst, P.P. Schot, andK. de Jong, 182. Utrecht, The Netherlands: Utrecht University.

van Eeten, M.J.G., D.P. Loucks, and E. Roe. 2002. Bringingactors together around large-scale water systems: Parti-cipatory modeling and other innovations. Knowledge, Tech-nology and Policy 14, no. 4: 94–108.

Vennix, J.A.M. 1996. Group Model Building: Facilitating TeamLearning Using System Dynamics. New York: Wiley.

Vennix, J.A.M., D. Anderson, and G. Richardson. 1997. Groupmodel-building, art and science. System Dynamics Review13, no. 2.

Videira, N., P. Antunes, R. Santos, and S. Gamito. 2003. Parti-cipatory modeling in environmental decision-making:The Ria Formosa natural park case study. Journal of Envir-onmental Assessment Policy and Management 5, no. 3:421–447.

Winz, I., and G. Brierley. 2007. The use of system dynamicssimulation in integrated water resources management. InProceedings of the 25th International Conference of theSystem Dynamics Society, System Dynamics Society,Boston, Massachusetts, July 29–August 2, 2008. Papers areavailable at http://www.systemdynamics.org/conferences/2007/index.htm.

Wondolleck, J.M., and S.L. Yaffee. 2000. Making CollaborationWork: Lessons from Innovation in Natural Resource Man-agement. Washington, D.C.: Island Press.

World Bank. 2003. Water Resources Sector Strategy: StrategicDirections for World Bank Engagement. Washington, D.C:World Bank.


Documents

Cooperative Modeling: Linking Science, Communication, and Ground Water Planning