The 33rd International Conference of the System Dynamics Society
Cambridge, Massachusetts, USA, July 19 - July 23, 2015
Toward Effective Water Diplomacy by Using System
Dynamics: Case Study
Mahdi Zarghami’”*, Mohammad A mir Rahmani*
' Department of Civil and Environmental Eng. Tufts University, Medford, MA 02155 USA
> Faculty of Civil Engineering, University of Tabriz, Tabriz, 51666, Iran
3 Sociotechnical Systems Research Center, Massachusetts Institute of Technology, Cambridge, 02142
USA
Emails: m. i@tufts.edu, m hmi@ tabrizu.ac.ir and Phone: +1 781 518 1417 and +98 41 3339 2549
Abstract
Rapid urbanization and climate change along with expanding agricultural and industrial water
rights are threatening all water resources, particularly lakes in arid regions such as Urmia Lake in
Iran. When these systems are healthy, they provide valuable ecological and socio-economic
benefits to current and future generations. However, many arid zone lakes are now subject to a
high risk of depletion, pollution, salt water intrusion, and aeolian transport. These processes may
have direct health consequences such as raising the risk of asthma, kidney stones, and some skin-
related diseases. These risks require robust solutions in the face of significant uncertainties and
potential conflicts among water stakeholders. The goal of this research is to propose sustainable
lake management alternatives by using a system dynamics (SD) approach. A simulation model to
compare the alternative rehabilitation options is developed focusing on feedback and the
extemalities of decisions made using the SD approach. The main aim of this modeling is to help
trust building among conflicting stakeholders, which is a key aspect of water diplomacy and
essential for effective water management. Results of the paper indicate that increasing irrigation
efficiency by 4% annually would be most effective plan in saving Urmia Lake. Zaab inter-basin
water transfer project could have more than 25% contribution in saving the lake, although that
would require some of the sensitive diplomacy and negotiation. Aras water transfers have low
feasibility in terms of water diplomacy. In addition, this study found that while decreasing 15%
of cultivated area over a period of six years would be more effective than cloud seeding,
reducing cropped area requires greater diplomacy than cloud seeding. This modelling approach
is a first step toward long term processes of decision support and water diplomacy in the Urmia
Lake basin and beyond.
Key words: System dynamics, Water diplomacy, Restoration, Urmia Lake.
1. Introduction
About 1.6 billion people, almost one quarter of the world's population, are facing economic
water shortages due to hydro-political fragmentation and conflicts among stakeholders. This
critical condition requires robust negotiations and solutions under conditions of great uncertainty.
As long as the ambiguity of these problems and potential solutions persist, it induces conflict
among stakeholders that results in further delay on any action to prevent loss of resources.
Water diplomacy is a process that involves joint-fact finding among stakeholders. The most
powerful aspect of this process is the shared information that it produces about the problem and
about cause-and-effect relationships between stakeholders’ actions. When it employs simulation
modeling, it provides a process to facilitate political negotiations and joint-understanding of the
problem. Simulation modelling can help decision makers to see the outcomes of different
policies in water management. Modelling interventions and feedback among system elements is
an essential part of identifying sustainable development solutions. The main advantage of this
modelling approach is that by simulating the effects of different policies on lake management
problems, it helps build trust among conflicting stakeholders.
An object-oriented system dynamics (SD) approach is a useful tool for modeling this type of
simulation problem (Forrester, 1973; Simonovic, 2009; Zarghami and Akbariyeh, 2012, Mirchi,
2013, Amir Rahmani and Zarghami, 2015). It allows us to conduct multi-scenario, multi-attribute
analyses that result in relative comparisons of many altemative management strategies over time
(Sahlke and Jacobson, 2005). A few lake management problems have been modeled using the
SD approach, especially relating to sustainability issues. For example, Guo et al. 2001 used this
approach to evaluate the environmental, social, and economic impacts of Chinese government
policy on the quality of water in Erhai Lake. The SD approach has also been applied to allocate
water in the Sun Juan Basin between Colorado and New Mexico (Ewers, 2005). Liu et al. (2008)
used SD to understand the effect of urban population increase and economic development on the
eutrophication of a lake.
The goal of this research is to generate sustainable lake management solutions to the water
control problems that affect Urmia Lake in Iran. This approach can help to integrate economic
prosperity, social equity and environmental health concems into a single framework. To achieve
this goal, a simulation model is first developed to model the socio-economic and ecological
processes for the Urmia Lake case study, and to assess policies to achieve its sustainability for
current and future generations. This approach can be generalized for other cases. The main
advantage of this research is that by simulating effect of different policies on the lake
management problem we may be able to avoid “fixes that fail.” For example, the Urmia Lake
Restoration Program has suggested nineteen solutions. These suggestions include hard and soft
solutions, such as inter-basin water transfer projects, cloud seeding, increasing agricultural water
efficiency and decreasing farmed land area. This paper simulates these restoration projects and
compares the relative effectiveness of them. This simulation model can help stakeholders reach
agreement and build trust through a process of water diplomacy modeling.
2. Problem Description: The Urmia Lake Case Study in Iran
Urmia Lake in northwestern Iran is the largest inland lake of the country and one of the largest
saline lakes in the world. The lake basin presents a highly influential and valuable aquatic
2
ecosystem. Approximately 1500 species of vascular plants, including the unique Artemia sp.,
inhabit the lake’s wetlands, representing 15% of the total number of flora species found in the
country. Because of its unique natural and ecological features, the lake and its surrounding
wetlands have been designated as a National Park, Ramsar Site, and a UNESCO Biosphere
Reserve (CIWP, 2008). The lake basin, as a socio-ecological region, has experienced extreme
water shortages in recent years due to poor water management and climate change (Alipour,
2006; Zarghami, 2011). Because of intense agricultural development and rapid urbanization, the
groundwater level in some parts of the basin has dropped by up to 16 meters. As a result, the
water level of the lake and its surrounding area are now below the critical level (Fig. 1). Wind-
blown salty dusts from dry areas of the lakebed can become a serious threat to the health of the
people residing in the area if the water inflow falls short of its minimum level of 3 billion cubic
meters to sustain the lake (A bbaspour and Nazaridoust, 2007).
Area (km?)
Fig. 1. Changes in the area of Urmia Lake (Sima and Tajrishi, 2013)
Given the severity of the situation, saving Urmia Lake is currently one of the top priorities of
national and several international organizations (including UNDP, UNEP, and JICA). However,
there is a lack of comprehensive research to evaluate the effects of alternative
adaptation/mitigation strategies on the lake. The goal of this research, therefore, is to provide a
technique to find the best, most robust strategies. This analysis can help build trust among the
stakeholders and then will provide a base to start water diplomacy framework. This research
deals with two important questions. The first set of questions concems the condition of Urmia
Lake in the near future, while the second evaluates effective strategies to survive the current
conditions. Theoretically, the first question leads to an argument about "the tragedy of the
commons," and the second toward an argument to prevent "fixes that fail." These two questions
are detailed in the following.
Questions about the problem at Urmia Lake: Will the lake dry up due to the overuse of its
water resources and climate change, as is happening in the Aral Sea? And if the lake does
disappear, what are the social, economic, and environmental impacts on the region? What would
be the effects on water quality and accessibility? Would this result in people from cities and
villages around the lake migrating to other cities? When might such problems start?
Questions about possible solutions: Which combinations of strategies would be most effective
and efficient in sustaining the basin? These options will be studied in two groups:
Soft solutions: water conservation, water market tools like price signals, water allocation
adjustments, and other organizational strategies.
Hard solutions: Reducing agricultural water use, preventing the construction of new dams,
inter- basin water transfer, and other infrastructural alterations.
3. Urmia Lake SD Model
The following steps were carried out to build this SD model: (i) describing the problem and its
uncertainties; (ii) defining key variables to develop the lake's model; (iii) preparing casual loop
diagrams and system archetypes; (iv) building a stock and flow model in the VENSIM software
environment; and (v) calibration and validation. Some of the needed data were drawn from
interviews and stakeholder meetings with local NGOs, national organizations and intemational
institutions such as UNDP. The author's close relationships with several associated institutions
and corporations made it relatively easy to collect the remaining data. To validate the
methodology, several tests including boundary adequacy, structure assessment, and dimensional
consistency were performed. The modeled lake level is compared with the observed data in the
last 20 years. Using the validated model, the hard and soft lake restoration strategies were
simulated for the next 20 years, considering various tradeoffs among the different objectives.
The main parts are the lake stock and water supply and extraction flows by the rivers and
groundwater as shown in Fig. 2.
‘Volume- Area “
( “a
i ‘| pfaDirect Rainfall
i
Ad ete
‘Outflow to Lake
hm
Drain from Lake to
Groundwater
\ Recharye Cot,
a Grow
Recharge Le r\ Gw a"
. £
i
\ \
Lake Area
— Evaporation Rate
J
Drainageto <2
Surface Water Si
Drainage-Gr
oundwater
Fig. 2 The main sub-model in the V ensim environment
The second sub-model is for demand forecasting. It aggregates water user’s demands including
the agriculture, industrial and domestic sectors. For domestic users, the main driver is the
population, which is modeled as shown in Fig. 3.
Pc st *
Al > —
| \ Demand
se
eae US
a “Domestic co
Disaficaion No Fd 1
Seprly | ———— Dom. nn
: Ins ee
Fig. 3 The main sub-model of the water demand
In the water supply side, the most important element is the rainfall-runoff model. It is presented
in Fig. 4. This sub-model includes the uncertainties in future rainfall due to cloud seeding and
also different climate change models.
A2Prec, Bi Prec, AIB Prec.
<Time>-
)
Cloud Seeding Cloud Seeding. ‘
Cost onloft
~_—
Climate Change
Scenatio
Cloud Seeding
Cloud Seeding PP a
Impact
—— Precipitation
—
ui a Runoft,
"Normal Year Plain
‘Area Runoff Coef,
te oie
<Precipitat {a Eg
Mountainous Area
Runoff of
re Mountainous Area
Runoff Coef,
‘Mountainous Area See
Runoff Coeff, <<
Fig. 4 The sub-models of rainfall and runoff
Finally the sub-model of conjunctive surface and groundwater use including percolating water
resources is presented in Fig. 5.
Controlled Water in
NonAgri Season
i= Time>
Noah Sa Season
Rate of Surface
Discharge in Agr Water Retum
Season
Ae sutce
aS sw Retum
Site Watt) =
Consumed Water
GW Retum
Supposed Lo Sy 60 Bhecions
Water Usage Surface Water ee
a SS eae
Rate of Surface
Water Consumption rata Deans
hae Rate of GW
Rate of Retum
Grounwater Usage
Supposed
Groundwater Usage
Fig. 5 The sub-model to explain the relations among water resources stocks and flow
Finally, simulation using the SyntheSim tool provides very valuable insight. It helps stakeholders
see the outcomes of using different policies for restoring the lake. The view of this visual result
is shown in Fig. 6. The initial model results presented here used stakeholder inputs from previous
studies and reports. The next stage of research will involve iterative validation of the model by a
group of regional and international experts in a water diplomacy simulation setting.
Uma Lake Volume
f ] 35,000. Milon Cubic Met
ae 1 [3 35,000. Milon Cubic Meter
‘Area RunoffCoei" Clitaie Change Scenario
— 17,500 Mion Cubic Meter
f
0
co «oC ]1 17500. Milon Cubie Meter JS
cc | Ea ee
"Normal Year Plain Area Runoff Coef." 0 Millon Cubic Meter a
— 0 Mien ce oer
o [1 ¢ 8 2 16 20 2% 28 92 36
cy 7 ‘Nim ono tite)
== c
bo ||
F
fo || = olume Age Area
Sal Cock 1 [__] 2» 30,000 400,000
‘Cloud Seating pact 4
——
j P
"Rect ‘Coek” ——| 20,000 300,000
— g H
rae j tH
10900 anno
+ ie 2 ais
—7 mab tn Tana
Cote vore ii
Fig. 6 “Control panel” provided by SyntheSim: Visual and dynamic representation of the Urmia
Lake volume affected by changes in the policies and parameters
In this study six restoration plans are considered. These policies, strategies and projects are
among the suggestions of the Regional Committee on Urmia Lake Basin, which is responsible
for saving the lake. Different action plans are defined as restoration scenarios to investigate if the
lake could be saved by various combinations of these plans. These projects and scenarios are
summarized in Table 1 and Table 2.
Plan 1- Increasing irrigation efficiency: Most academics and engineers agree that agricultural
water use efficiency is low in Iran. Hence, increasing water efficiency is considered as a solution
for water scarcity problems. It is assumed in a ten-year plan that water efficiency could be raised
to 70% in the basin (i.e., a 4% annual increase in this parameter). Changing crop pattems and
improving irrigation systems could effectively increase water use efficiency.
Plan 2- Reducing agricultural area: Rapid growth in farm land area in the basin is a major
factor for the Urmia Lake disaster. Therefore, stopping its growth and decreasing area are
followed in plan 2. Since the area of decrease is a major source debate among stakeholders, an
alarm variable is defined in the model. This variable represents the degree of vulnerability of the
lake and of groundwater fall. The area of decrease is estimated by the alarm variable, but it is
assumed to have maximum reduction about 5% considering the social resistance and sensitivity
regarding this plan.
Plan 3- Cloud seeding: Cloud seeding to increase precipitation over an area could make an
effective contribution toward reducing water stress in a basin. Consequently, adding some water
to the basin by cloud seeding could help in lake restoration. Cloud seeding impact differs based
on the seeding method, climate of the region, and type of clouds. But it is assumed here that
cloud seeding could enhance annual precipitation by up to 7%. This effectiveness value is the
average of numbers we found in the literature (Curic et al. 2007, Silverman 201, Acharya et al.
2011, DeFelice et al. 2014)
Plans 4 and 5- Water Transfer: The other way of adding water to a basin is via inter-basin
water transfer projects. Two main water transfer projects considered here are from the Zaab and
Aras basins. The Zaab basin could transfer 600 MCM annually to Urmia Lake basin to supply
increasing water demands. Also there are plans to transfer 140 MCM annually from Aras basin
north of Urmia Lake, though this is an international river basin tributary to the Caspian Sea that
does not have clear water allocation agreements.
Table 1. Name of Plans and their action
Plan Action Horizon Effect
P.1 Increasing irrigation efficiency 2015-2025 4% annually
P.2 Reducing agricultural area based on drought intensity 0%-5% annually
PS. Cloud seeding 2015-2030 7% annually
P.4 Transfer from Zaab 2015-2030 600 MCM/Y
P.5 Transfer from Aras 2015-2030 140 MCM/Y
In addition to the volume of the lake that serves best for measuring restoration success, two other
indicators were selected to check if the plans lead the basin toward sustainability. Relative water stress
indicator (RWSI) computes the demands on available water resources in a basin. RWSI formulation is
presented in Equation (1).
RWSI = oF (1)
which DIA is total demand (MCM) in a basin including domestic, industrial and agricultural water needs
and Q total available surface and ground water (MCM). RWSI>0.4 for a basin indicates a highly stressed
and critical condition. Second, groundwater dependency (GD) expresses the degree of groundwater
participation in water supply in a basin. This indicator can be computed by Equation (2):
CD = totalgroundwater abstraction _ (2)
total surface and ground water supplies
Favorable value for GD is less than 0.25. Represented GD for basins with high dependency to ground
water is more than 0.5.
Modelling results: Fig. 7 illustrates lake level changes under each plan. Plan1 follows policy of
increasing water efficiency by 4% annually. Hence, it is supposed to have 70% water efficiency by 2025
in the basin. It is seen improving this parameter could not save the lake on its own, because a direct
feedback of this policy is increasing land area and crop density by farmers. It is estimated RWSI and GD
are almost 0.85 and 0.34 in the basin under this plan. Hence, implementing only this plan neither restores
the lake nor leads the basin to sustainability because of the socio-economic response of the basin.
Decreasing farmed land area which is Plan 2 is seen just saves the lake to minimum ecological level
(MEL). This plan needs to lower area of cultivation 20% in a 5-year duration to lead the lake to MEL in
2020. RWSI and GD are estimated to be almost 0.54 and 0.43 in the basin under this plan. Certain crucial
socio-economic consequences of this plan in the region would be considerable.
It is seen in Fig. 7 adding extra water to the basin by cloud seeding can save the lake to MEL on its own.
It is assumed this project raises precipitation 15% in the basin, though the high scientific and technical
uncertainties of this estimate must be acknowledged. Cloud seeding project improves GD to 0.28, but it
makes RWSI worse in the basin.
Inter-basin transfer projects of Zaab and Aras are supposed to have considerable favor for restoration of
Umnia Lake. Zaab basin is supposed to transfer 600 MCM under Plan 4 and Aras 140 MCM under Plan 5
to Ummia Lake basin. Based on Fig. 7 it appears these projects can help the lake from drying up and but
save the lake. It must be noted that these international plans would require additional levels of water
negotiation and diplomacy.
1276.4 1276.4
1275.9 1275.9
125A 12754
a 1274.9 7G 1274.9
= 12744 = 12744
& 1273.9 | § 1a73.9 [>
& 1273.4 pra PS
a a . ee
1272.9 1272.9 senate
g ‘geri
BL Bn
SeSSARANGSRRR ARS SeSSSS ANS S98 8828
RRSRRRRRRRRRRRRA RRRRRERRRRRRARRRS
—>- Increasing Agricultural Efficieny - - Min. Ecol. Lev. ~« Less AgricultureArea ~ - Min. Ecol. Lev.
1276.4 1276.4
1275.9 1275.9
12754 , 12754
@ 1274.9 @ 12749
é & 12744
: § 1273.94 00
§ 1273.4
fa ia pai os Nees
g § 12729 Le sia
4 Ama °
a
SSR eaRANRARSHRAS
RRSRRSRRRRRRRRRRKA
* ZaabTransfer ~~ Min. Ecol. Lev.
1276.4
1275.9
1278.4
7 1274.9
= 1274.4
5 og See
i 1273.4
F 12729 rg ite SG sw
4 1272.4
2ERZIgn NASH SRRAB
PPP EEEEELELEELEE
-s-ArasTransfer ~~ Min. Ecol. Lev.
Fig. 7 Simulation of the individual effect of each plan by Urmia SD model
10
A combination of these plans is likely to be implemented to save Urmia Lake. Hence, assuming all of the
plans are in action, estimating the contribution of each plan supports the objectives of this paper. In the
scenario of implementing all 5 plans, RWSI is estimated to be almost 0.6 and GD almost 0.4. Although
RWSI indicates the basin is still highly stressed based on international standards, it is below average
RWSI in Iran, which could make it acceptable. When all plans are implemented, the dynamics of the
basin are totally different from the situation in which each plan is applied alone. In this scenario there is
more water in the basin because of inter-basin transfers and cloud seeding, and as a result, smaller area
needs to be taken out of farming. The percent contribution of each plan is plotted in Fig. 8.
Increasing
Water Efficieny;
32%
Zaab Transfer; Aras Transfer,
2% Th
Cloud Seeding;
15%
Fig. 8 Contribution of each plan in restoring Urmia Lake to its ecological level (% )
Based on the results of this paper, implementing all plans could cause Urmia Lake to be restored by 2022
even though some plans are not completed by that time. It is seen increasing water efficiency has the most
important share in saving Urmia Lake. In the second place, Zaab Transfer project has ~ 27% effect in
raising lake level. Stopping 15% of cultivated area from farming (in restoration period) has 19%
contribution in reaching to minimum ecological elevation. Having 7% increase in annual precipitation by
cloud seeding has a contribution about 15% in saving the Lake. Based on these results shown in Fig. 8,
Aras transfer project has an impact less than 7% in the basin.
4. Conclusions
This paper presents the first results of Urmia SD model to support a process of water diplomacy. The
authors drew upon stakeholder reports to generate this model, and plan to solicit comments from
stakeholders and experts on the Urmia Lake to update the model. Based on these initial results, increasing
water efficiency in a ten-year period is the most effective plan for restoring Urmia Lake, and may also be
the most feasible among stakeholders. The Zaab water transfer has the second place among plans for
11
restoring the lake to its minimum ecological level, but it would involve high levels of water diplomacy.
Also, it is seen that decreasing agricultural area to keep it 85% of its existing area within a 6-year period
(2016-2021) could make a contribution of almost 20% to lake level restoration. Although the modeling
approach of this research differs from that of the Regional Council on Lake Urmia Basin Management,
(2012), the results are close and affirming.
Using this system dynamic model can help the stakeholders see the effects of different policies. It can
help build trust among them which is vital for long-term water diplomacy. However, since restoring the
lake is a complex problem, more studies are needed. One important job that is in progress is comparing
alternative plans by multi-criteria analysis. Finding near-optimal combinations of cultivated land areas to
restore the lake and also improve agricultural sustainability is vital. Conflict resolution mechanisms and
experiments with stakeholder groups need to be added to the model in future research.
Acknowledgements
The support of University of Tabriz for the sabbatical leave of the first author at Tufts
University/ MIT is appreciated. The authors are also thankful for comments and contributions of
Prof. Shafiqul Islam (Tufts University), Prof. James Wescoat (MIT), M. Mahdi Hashemian
(MIT), Terrence Smith (Tufts University), and John Ikeda (World Bank) on the research idea.
References
Abbaspour, M., and Nazaridoust, A. 2007. Determination of environmental water requirements of Lake Urmia, Iran:
an ecological approach. International Journal of Environmental Studies, 64(2):161-169.
Acharya, A., Piechota, T.C., Stephen, H., Tootle, G. 2011. Modeled streamflow response under cloud seeding in the
North Platte River watershed, Journal of Hydrology, 409(1-2): 305-314.
Alipour, S. 2006. Hydrogeochemistry of seasonal variation of Urmia Salt Lake, Iran, Saline Systems, 2(9),
doi:10.1186/1746-1448-2-9.
Amir Rahmani, M., Zarghami, M. 2015. The use of statistical weather generator, hybrid data driven and system
dynamics models for water resources management under climate change. Joumal of Environmental
Informatics 25(1): 23-35.
Curic, M., Janc, D., Vuckovic, V. 2007. Cloud seeding impact on precipitation as revealed by cloud-resolving
mesoscale model. Meteorology and Atmospheric Physics, 95(3-4): 179-193.
CIWP: Conservation of Iranian Wetlands Project. 2008. Integrated management plan for Lake Urmia, Tehran:
Department of Environment.
DeFelice, T.P., Golden, J. Griffith, J., Woodley, W., Rosenfeld, D., Breed, D., Solak, M., Boe, B. 2014. Extra area
effects of cloud seeding — An updated assessment, Atmospheric Research, 135- 136: 193-203.
Ewers, M. 2005. Combining hydrology and economics in a system dynamics deling water
for the San Juan Basin. Proc., 23 International Conference of the System Dynamics Society, July 17-21,
Boston.
Forrester, J.W. 1973. World Dynamics, 2nd ed. Wright Allen Press, Cambridge, Massachusetts.
Guo, H.C., Liu L., Huang G.H., Fuller G.A., Zou R., and Yin Y.Y. 2001. A system dynamics approach for regional
environmental planning and management: a study for the Lake Erhai Basin. Journal of Environmental
Management, 61(1): 93-111.
Liu, Y., Guo H.C., Yajuan Y.U., Dai Y.L., Zhou F. 2008. Ecological- i Jeling as a tool for hed
management: a case study of Lake Qionghai watershed, China. Limnologica - Ecology and Management of
Inland Waters, 38(2): 89-104.
Mirchi, Ali, 2013. System dynamics modeling as a quantitative-qualitative framework for sustainable water
resources management: Insights for water quality policy in the great lakes region. Dissertation, Michigan
Technological University.
Sahlke, G., and Jacobson J. 2005. System dynamics modeling of transboundary system: The Bear River basin
model, Groundwater, 43(5): 722-730.
Regional Council on Lake Urmia Basin Management, 2012. Drought Risk Management Plan for Lake Urmia Basin,.
Technical Report.
Silverman, B.A. 2010. An evaluation of eleven ional cloud seeding in the sheds of the Sierra
Nevada Mountains, Atmospheric Research, 97(4): 526-539.
Sima, S., and Tajrishy, M. 2013. Using satellite data to extract volume-area-elevation relationships for Urmia Lake,
Iran. Journal of Great Lakes Research, 39(1): 90-99.
Simonovic, S.P. 2009. Managing Water Resources, Methods and Tools for a Systems Approach, Paris: UNESCO
Publishing / Earthscan.
Zarghami, M. 2011. Effective watershed management; Case study of Urmia Lake, Iran, Lake and Reservoir
Management, 27(1): 87-94.
Zarghami, M. and Akbariyeh, S. 2012. System dynamics modeling for complex urban water systems; A pplication to
the city of Tabriz, Iran, Resources, Conservation and Recycling, 60: 99-106.
