[Link].

com/scientificreports

OPEN Evaluation of cooperative

and non‑cooperative game
theoretic approaches for water
allocation of transboundary rivers
Fahimeh Mirzaei‑Nodoushan1, Omid Bozorg‑Haddad1* & Hugo A. Loáiciga2
Efficient water allocation in a transboundary river basin is a complex issue in water resources
management. This work develops a framework for the allocation of transboundary river water
between the countries located in the river basin to evaluate the characteristics of allocation
approaches. The allocation of river water is obtained based on initial-water conditions, cooperative,
and non-cooperative game-theoretic approaches. The initial-conditions water allocation approach
assigns 34, 40, and 26% of the Harirud River flow to Afghanistan, Iran, and Turkmenistan, respectively.
The game-theoretic cooperative approach assigns 36, 42, and 22% of the river flow to Afghanistan,
Iran, and Turkmenistan, respectively. The non-cooperative game-theoretic approach establishes
that the most stable water allocation was 42, 38, and 20% of the Harirud River flow for Afghanistan,
Iran, and Turkmenistan, respectively. Human and agricultural water-stress criteria are used to
evaluate the water allocations in the Harirud River basin. The criterion of human water stress has
the largest influence in Iran, and the criterion of agricultural water stress has the smallest influence
in Afghanistan. This work’s results indicate the initial-conditions water allocation approach favors
Turkmenistan, whereas the cooperative and the non-cooperative game-theoretic approaches favors
Iran and Afghanistan, respectively. The results show that the priorities of each country governs water
allocation, and cooperation is shown to be necessary to achieve sustainable development.

Water management in transboundary river basins involves technical and scientific assessments, and legal, eco-
nomic, cultural, social, and historical factors that cross political and cultural boundaries. Transboundary river
management can improve food security, economic growth, adaptation to climate change, international trade,
public health, and international cooperation (Mirzaei-Nodoushan et al.1). Management of transboundary river
basins requires adaptive management which means the countries can adapt to changes and stresses of water
resources and reduce their effect on sustainability (­ Wolf2).
Sustainable development is an effort to balance between economic evolutions, environmental issues, and
social and welfare needs of a region in the present without damaging the future generations’ resources (Karimian
et al.3). The challenges of the water crisis increase in shared water regions, and water managers must cooperate
to manage water in a beneficial manner. Determining the benefits from the cooperation between countries and
how they are distributed among the stakeholders is a nontrivial task in transboundary river basin management
­(Wolf2). To this end, it is necessary to increase the institutional capacity so that beneficial water management
and water use can be established at all levels of society, and to ensure that governance organizations become
active and effective in decision-making.
The allocation of water resources among transboundary stakeholders has been approached from various
perspectives (Bozorg-Haddad et al.4). The most common solutions for water allocation are optimization methods
and game theory approaches (Akbari-Alashti et al.5). As an example, Kazemi et al.6 reported a multi-objective
optimization model for the allocation of water resources in the Sefidrood River Basin, Iran. The objectives of
their model were to maximize revenue and minimize the Gini index. The results showed that the Gini index
increases (an increase of injustice) with the increasing number of dams.
The genesis of game theory is traced to von Neumann and ­Morgenstern7, who provided a mathematical
method for the analysis of human behavior in strategic decision-making by competitive or cooperative entities.

1
Department of Irrigation & Reclamation Engineering, Faculty of Agricultural Engineering & Technology, College
of Agriculture & Natural Resources, University of Tehran, Karaj 3158777871, Iran. 2Department of Geography,
University of California, Santa Barbara, CA 93106, USA. *email: OBHaddad@[Link]

Scientific Reports | (2022) 12:3991 | [Link] 1

Vol.:(0123456789)
 [Link]/scientificreports/

­ ash8 revolutionized game theory with the introduction of the Nash equilibrium for n-player games. The Nash
N
equilibrium defines the solution of non-cooperative games. It occurs when each player knows the equilibrium
strategies of the other players and no player has anything to gain by changing only his or her strategy (Holt
and ­Roth9). Many decision-making situations involving the allocation of scarce resources under cooperative
or non-cooperative behaviors can be solved with game theory (Harrison and L ­ ist10), as done in this work for
the allocation of a transboundary river’s water. Degefu et al.11 proposed a framework in which the theory of
bargaining was combined with the allocation of resources and bankruptcy games to allocate water and improve
welfare in the transboundary river basin. The Euphrates River was selected as the study area that encompasses
parts of Turkey, Syria, and Iraq, and a framework was created for the economic redistribution of water and
maximizing overall welfare. Arfanuzzaman and Abu S­ yed12 examined the relationship between ecosystem and
water needs in the Teesta River Basin between India and Bangladesh using game theory. Their results showed
that the zero-sum game perspective can be used to develop transboundary water policies. Zeng et al.13 applied
a cooperative game theory method to solve transboundary water problems between two cities in the Guanting
basin of northern China. The results demonstrated that full cooperation leads to the greatest total benefit to each
stakeholder. Qin et al.14 proposed an integrated decision support framework combining multi-criteria decision
making, bankruptcy theory, and the power index in the Dongjiang River Basin in China. The results showed
that the proposed method is reliable for simultaneously considering the criteria of fairness and sustainability in
the management of transboundary rivers.
The HBV model is a lumped conceptual catchment model that has relatively few model parameters and
minimal forcing input requirements, usually the daily temperature and the daily precipitation (Bergström and
Göran15). HBV stands for the Swedish title Hydrologiska Byråns Vattenbalansavdelning which was developed
in 1976. The model is a popular rainfall-runoff model in simulating gauged and ungauged basins (Yang et al.16)
that has been applied in different climate regions (Seibert and B ­ even17; Samuel et al.18; Pool et al.19).
Cooperative management of transboundary rivers contributes to sustainable water management and improves
the chances of rendering benefits for all riparian countries (Mirzaei-Nodoushan et al.1). Yet, riparian countries,
especially upstream ones in transboundary basins, are sometimes hesitant to be cooperative, and they tend to
pursue self-serving use of river water. In this context, it seems necessary to consider strategies for water alloca-
tion according to realistic conditions within transboundary basins and to examine the effect of the strategies on
transboundary countries. Past studies of water allocation lack a consistent framework to assess water allocation
in transboundary basins.
This paper presents and analyses three approaches for transboundary water allocation that is displaying
different decision-making of stakeholders. This study intends to show two types of common decision-making
(behavior) of countries in water allocation and show the effect of these behaviors in meeting the water needs of
countries that are not presented before in past studies. First is the allocation of water among competing users
according to an initial-conditions water allocation of water that is made based on geographic location, economic
development, and historical water use. The second approach involves game theory, wherein several coalitions
of water users (or players in the game-theory jargon) are formed in which the players have specific functions.
Coalition revenues are distributed among the players of a coalition based on inter-player cooperation and their
agricultural, urban, industrial, and environmental water requirements. The third approach for transboundary
water allocation is based on non-cooperative game theory. Finally, three approaches are evaluated based on the
riparian countries’ characteristics.

Methods
The genetic algorithm (GA). The GA is a pioneering evolutionary algorithm for optimization ­(Holland20).
The GA and its variants mimic evolutionary processes in nature (mutation, cross-over, survival of the fittest) to
improve an initially, randomly, generated population of possible solutions iteratively until it converges to a near-
global solution. The reader is referred to Bozorg-Haddad et al.21 for a description of the GA and some of its vari-
ants with codes available for their programming.

Mathematical programming model. This section presents a general mathematical programming model
for a transboundary river system that serves urban, agricultural, and industrial water demands.

Objective functions. Two objective functions were implemented in this study. One maximizes the water supply
reliability and was applied to obtain an initial-conditions water allocation. The second objective function maxi-
mizes the agricultural revenue by applying cooperative and non-cooperative game theory. Reliability measured
the probability of a system performing its function satisfactorily. Volumetric reliability, specifically, was equal to
the volume of water supply over the volume of water demand during the operational period. The equation for
volumetric reliability (αv ) is as follows:
N
AW t
αv = t=1N
, (1)
t=1 Det

where AW t = volume of water supplied at time t, and Det = volume of water demand at time t. AW t and Det
were calculated as follows:
Ik
K 

AW t = (XAi,k,t + XU i,k,t + XDi,k,t ), (2)
k=1 i=1

Scientific Reports | (2022) 12:3991 | [Link] 2

Vol:.(1234567890)
[Link]/scientificreports/

for all t , where XAi,k,t , XU i,k,t , and XDi,k,t denoted the volumes of water used by the agricultural, urban, indus-
trial sectors at location i in country k and time t, respectively. The agricultural water use at location i , country
k , and time t are given by:
Jk

XAi,k,t = Ar i,j,k,t × Iri,j,k , (3)
j=1

for all i , k , t , where Ar i,j,k,t , Iri,j,k, = area of crop j at location i in country k and time t , and consumptive (water)
use of crop j (volume of water per area) at location i in country k , respectively. The volume of water demand at
time t is calculated with Eq. (4):
Ik
K 

(4)
 
Det = DeAi,k,t + DeU i,k,t + DeDi,k,t ,
k=1 i=1

for all t, where DeAi,k,t , DeU i,k,t , DeDi,k,t denote the agricultural, urban, and industrial sectors’ water demands
or needs, respectively, at location i, country k, and time t.
The agricultural revenue is calculated as follows:

Bagr = Pr j,k × Ar i,j,k,t × Yd i,j,k ,
(5)
t k i j

where Bagr , Pr j,k , and Yd i,j,k = the agricultural revenue, price (say, in $/ton of produce) of crop j in country k, and
the yield per unit area (say, in ton/hectare) of crop j at location i in country k, respectively. The revenue to country
k = r, 1 ≤ r ≤ K , is calculated with Eq. (5) by summing over t, i, j while setting the country index constant and
equal to r . The decision variables of the mathematical programing problem are the areas Ar i,j,k,t and the urban
( XU i,k,t ) and ( XDi,k,t ) industrial water uses.

Constraints. Water balance at each location i of country k and time t :

Qout k,i,t = Qink,i,t − XAi,k,t − XU i,k,t − XDi,k,t , (6)
for all i, k, t , where Qout k,i,t = the volume of water leaving location i in country k at time t, Qink,i,t = the volume
of water entering location i in country k at time t.
The water used by the agricultural, urban, and industrial sectors may not exceed their water demands. The
constraints on water use are as follows:
XAi,k,t ≤ DeAi,k,t , (7)

XU i,k,t ≤ DeU i,k,t , (8)

XDi,k,t ≤ DeDi,k,t , (9)

for all i, k, t . The nonnegativity constraints on water use and outflow are as follows:
XAi,k,t , XU i,k,t , XDi,k,t ≥ 0, (10)

Qout k,i,t ≥ 0, (11)

for all i, k, t .

Game theory. Two types of game theory for decision making, cooperative and non-cooperative game the-
ory, are described in this section. The countries sharing the transboundary river are modeled as players of a
competitive game vying to maximize their payoffs, which in this case means maximizing the reliability of their
water supply and their agricultural revenues.

Cooperative decision making. Each player is in this instance is primarily interested in the magnitude of his or
her revenue from participating in a cooperative game rather than in increasing the total income of the coopera-
tive game compared to what could be obtained by participating in a non-cooperative game. In a cooperative
game players interact with each other before the game and they may sign contracts with each other and form
alliances. The revenue from all possible strategies is expressed in the form of characteristic functions, such as
revenue received from a game. A fair revenue sharing by cooperative players can be obtained with the Shapley
value method ­(Shapley22).

Non‑cooperative decision making. It is assumed in non-cooperative games that each player acts rationally in
choosing his or her strategy. The choice of strategies in this type of game aims at maximizing the individual
interests of each player. The concept of the Nash equilibrium is central to non-cooperative game theory. Nash
equilibrium assumes that each player can only have one strategy during the game, and each player considers the
strategies of the other players when making decisions. The Nash equilibrium may fail to accurately predict the

Scientific Reports | (2022) 12:3991 | [Link] 3

Vol.:(0123456789)
 [Link]/scientificreports/

Rainfall-runoff simulation

Determining transboundary Defining of constraints and

river evaluation criteria objective functions

Preparing mathematical
programming model

Initial-conditions water
allocation

Determining game theory

players and strategies

Water allocation with the Water allocation with the

cooperative game approach non-cooperative game approach

Evaluating approaches

Figure 1.  Methodology flowchart.

interaction between players in reality. For this reason, other methods are used to improve the modeling of non-
cooperative games in which each player attempts to anticipate the decisions of his or her rivals when choosing
a strategy.

Criteria for evaluating transboundary water management. Various political, social, economic, and
environmental criteria are defined to determine the status of transboundary water resources. The United Nations
Environment Programme (UNEP) has set out specific criteria for assessing the management of transboundary
basins in the form of a codified approach (Hooper and ­Lloyd23). This study considers the human water stress and
the agricultural water stress criteria, which are defined next.

Human water‑stress criterion. The human water stress proposed by the United Nations is measured by the
amount of water available per person per year. This criterion was obtained by dividing the amount of available
water by the population served:
AW
HI = , (12)
Tpop

where HI = human water-stress criterion, AW = amount of available water and Tpop = population served. The
lower the value of this criterion, the higher the human water stress.

Agricultural water stress criterion. The agricultural water-stress criterion is equal to the amount of water avail-
able in a basin divided by the area of agricultural land:
AW
AI = , (13)
Ar
where AI, AW, and Ar denote the agricultural water stress index, the volume of water available and the area of
agricultural land respectively. Thus, the lower the value of this criterion, the higher the agricultural water stress.
This paper’s methodology is displayed as a flowchart in Fig. 1, where it is seen that the river runoff was first
obtained with the HBV rainfall-runoff model. The mathematical model given by Eqs. (1)–(11) was developed.
The water allocation was obtained with the initial-conditions approach, and the cooperative and non-cooperative
game theory approaches were then calculated. Lastly, the water allocations were evaluated with the human water-
stress and the agricultural water-stress criteria.

Scientific Reports | (2022) 12:3991 | [Link] 4

Vol:.(1234567890)
[Link]/scientificreports/

Figure 2.  Geographical location of the Harirud river basin (ArcGIS, version: 10.5.0.6491, site: www.​esri.​com).

Study area. The Harirud River is a transboundary river in eastern Iran that flows through Afghanistan, Iran,
and Turkmenistan and has been a source of water for all three countries since ancient times. Much of the river’s
water flow, which originates in the highlands of Afghanistan, is controlled by the latter country through dams,
leading to a significant reduction in water flow in the downstream countries (see Fig. 2).

Ethics approval. All authors accept all ethical approvals.

Consent to participate. All authors consent to participate.

Consent for publication. All authors consent to publish.

Results
There was lack of data in the Harirud Basin outside Iran and minimal data sharing between the transbound-
ary countries. It was therefore necessary to apply rainfall-runoff simulation to make up for the data shortages.
Rainfall-runoff simulations for the Harirud River basin were performed using the HBV conceptual model. Data
on precipitation, temperature, potential evapotranspiration, and river flow for a 10-year (1969 through 1978)
were implemented for model calibration. The observational time series of monthly river flow corresponded
to the Robat Akhund gaging station in Afghanistan. The calibrated HBV model simulated monthly river flow
from 1979 through 2014. The correlation coefficient (R2) of the simulated flows equaled 0.821, which indicated
acceptable predictive skill or accuracy. The river water was allocated in this study based on the water demands
within the Harirud River basin for the period 2000–2014.

Initial‑conditions water allocation. The initial-conditions water allocation was calculated with the GA
by maximizing the reliability of water supply. Water is allocated to the countries according to their water needs,
and there are no constraints concerning cooperation to maximize the sum of agricultural revenue nor on non-
cooperating countries and their behavior effects with respect to received water. This means the only constraints
are those introduced in the mathematical programming model, and the objective function is to maximize the
reliability of water supply. The average annual volumes of water supply and demand for each country calculated
with the initial-conditions allocation are depicted in Fig. 3, where it is that the total water demands of countries

Scientific Reports | (2022) 12:3991 | [Link] 5

Vol.:(0123456789)
 [Link]/scientificreports/

1400

1200

Water Volume (MCM)

1000

800
Supply
600
Demand
400

200

0
Afghanistan Iran Turkmenistan
Country

Figure 3.  The average annual volume of water supply and demand for each country calculated with the initial-
conditions water allocation approach. (MCM = ­106 ­m3).

Country Shapley value (M$ = ­106$)

Afghanistan 13.23
Iran 27.63
Turkmenistan 6.37

Table 1.  Shapley value of countries.

Country Gately index

Afghanistan 1.30
Iran 1.14
Turkmenistan 9.39

Table 2.  Gately index of countries.

are much larger than the available water. The water is allocated to the countries according to their demands so
that the percentages of river flow allocated to Afghanistan, Iran, and Turkmenistan are on average equal to 34,
40, and 26, respectively.

Water allocation with the cooperative game approach. The mathematical programming problem of
water allocation maximizes the agricultural revenue. The priority of water withdrawal by the upstream country
(Afghanistan) is considered because of its geographically advantageous position. The players in the coalition
of cooperators must agree on revenue sharing. This sharing was obtained with the Shapley value method that
establishes the relative contribution of each player to the revenue of the coalition (­ Shapley22). The contributions
are listed in Table 1, where it is seen that Iran’s contribution to the coalition’s revenue is the highest and Turk-
menistan’s contribution is the lowest. Iran earns more agricultural revenue due to its larger water demand and
higher crop yields.
A key issue in a cooperative game concerns the stability of a coalition, which can be evaluated with the
Gately index. This index measures a player’s ability to harm other players if it leaves the coalition (­ Gately24). The
calculated Gately index is shown in Table 2, where it was seen that Turkmenistan’s index is higher than those
of the other countries. This implies that Turkmenistan is the most eager to leave the coalition. The Gately index
establishes that Turkmenistan would benefit more if it only partners with Afghanistan or Iran. On the other
hand, the value of this index for Iran is less than those of the other countries, which means that it would cause
more harm to itself than to the other countries if it leaves the coalition.
The water allocations obtained with the cooperative approach are displayed in Fig. 4. The water allocations
to Afghanistan, Iran, and Turkmenistan equalled 36%, 42%, and 22%, respectively, of the river flow. This implies
that Afghanistan’s and Iran’s shares of river water increases and Turkmenistan’s share decreases in comparison to
the initial-conditions water allocation. This approach dictates that a country with higher crop yields and wider
arable land acquires a larger share of the available water. The upstream power of water withdrawal by Afghanistan
means that this country enjoys a significant volume of water, despite having fewer inhabitants and less arable
land than Iran. Also, Turkmenistan’s water share decreases because of its lower crop yields and having less arable
land than Iran, which affects the objective function maximizing agricultural revenue.

Scientific Reports | (2022) 12:3991 | [Link] 6

Vol:.(1234567890)
[Link]/scientificreports/

1400

1200

Water Volume (MCM)

1000

800
Supply
600
Demand
400

200

0
Afghanistan Iran Turkmenistan
Country

Figure 4.  Average annual volume of water supply and demand calculated with the non-cooperative allocation
approach (MCM = ­106 ­m3).

Turkmenistan
Afghanistan Iran Compromise Opposition
Compromise 44.71 63.60 12.34 45.51 65.67 12.98
Compromise
Opposition 47.03 63.76 12.68 46.29 64.80 11.74
Compromise 46.89 64.70 13.52 49.93 66.15 11.79
Opposition
Opposition 48.10 60.91 13.09 48.30 61.15 10.08

Table 3.  Country payoff matrix corresponding to the countries’ strategies (annual revenue M$ = ­106$).
Significance values are given in bold.

Water allocation with the non‑cooperative game approach. Each country devises strategies for
interacting with the other countries. Afghanistan occupies the upstream portion of the Harirud Basin, and it
may pursue a compromise strategy and allow larger flow to the downstream countries, resulting in larger rev-
enue for Iran and Turkmenistan. Afghanistan, on the other hand, may use more water and reduce the down-
stream revenue. The downstream countries have two strategies: (1) they can choose a strategy of compromise
with Afghanistan’s strategy seeking to improve basin management, or (2) they can adopt a strategy of opposition
to Afghanistan’s strategy and seek to undermine it.
The formulation of the non-cooperative game is adjusted by incentives from the downstream countries to
the upstream county, and the optimization is implemented to maximize agricultural revenue. The incentive is
structured such that Iran and Turkmenistan would give Afghanistan $1.5 million and $0.5 million, respectively,
for each additional million cubic meters of water that Afghanistan releases to the downstream countries.
The combinations of strategies of the three countries are listed in the payoff matrix of Table 3. The values
of the payoff matrix represent the annual revenue accruing from the use of the Harirud River water received
by Afghanistan, Iran, and Turkmenistan. The payoff to each country specified by the payoff matrix depend on
its strategies (compromise or opposition) and the combinations of the other countries’ strategies. The strategy
corresponding to each situation yielding the highest payoff constitutes the Nash equilibrium of that country
corresponding to that situation. Examining the payoff matrix reveals that the only situation in which the three
countries reach a state of equilibrium when Afghanistan has an oppositional strategy and Iran and Afghanistan
have compromising strategies.
The annual water allocation calculated with the non-cooperative approach when Afghanistan is opposi-
tional and Iran–Turkmenistan compromise is depicted in Fig. 5. In this situation the Nash equilibrium is of a
non-cooperative nature. The water allocations to Afghanistan, Iran, and Turkmenistan represent 42, 38, and
20% of the river flow, respectively. In this instance Afghanistan’s share of river water is the largest, and Iran’s
and Turkmenistan’s shares are the lowest in comparison to the other allocations. The results obtained in this
approach are closer to the current state of water allocation in this basin. Yet, Iran has a lower share of water, and
Turkmenistan has a larger share of water in the current state of water allocation than the shares prevailing in the
non-cooperative allocation approach. The water released downstream by Afghanistan featuring an oppositional
stance is divided equally between Iran and Turkmenistan.

Analyzing the results of the calculated allocations. The combined annual revenues of the coun-
tries corresponding to the three allocation approaches are depicted in Fig. 6. The largest agricultural combined
revenue for Afghanistan, Iran, and Turkmenistan is obtained with the cooperative approach, and the initial-
conditions water allocation produces the smallest revenue. Turkmenistan has a larger share of water and less
value-added agriculture according to the initial-conditions water allocation approach. In this instance the total
annual agricultural revenue of the countries is reduced. Iran’s share of river water increases when there is coop-

Scientific Reports | (2022) 12:3991 | [Link] 7

Vol.:(0123456789)
 [Link]/scientificreports/

1400

1200

Water Volume (MCM)

1000

800
Supply
600
Demand
400

200

0
Afghanistan Iran Turkmenistan
Country

Figure 5.  Average annual volume of water supply and demand calculated with the non-cooperative allocation
approach (MCM = ­106 ­m3).

124
Agricultural Revenue (M$)

122

120

118

116

114

112

110

Initial-Conditions Cooperative Approach Non-Cooperative Approach

Figure 6.  Total annual revenue calculated with the three allocation approaches (M$ = ­106$).

70.00

60.00
Agricultural Revenue (M$)

50.00

40.00

30.00

20.00

10.00

0.00
Afghanistan Iran Turkmenistan
Country
Initial-Conditions Cooperative Approach Non-Cooperative Approach

Figure 7.  Annual agricultural revenue of each country calculated with the three allocation approaches
(M$ = ­106$).

erative sharing compared to the other two approaches (i.e., initial-conditions and non-cooperative), and agricul-
tural revenues are largest due to the high value-added of Iran’s agriculture compared to those of other countries.
The individual countries’ revenues are displayed in Fig. 7, where it is seen that the largest revenue for Afghani-
stan, Iran, and Turkmenistan corresponds to the non-cooperative, cooperative, and initial-conditions alloca-
tions, respectively. It pays for Afghanistan to be oppositional because the Nash equilibrium is achieved when its
strategy is oppositional and the other two countries compromise. Therefore, Afghanistan does not release excess
water downstream, but the two downstream countries compromise and provide incentives to Afghanistan to
release water. The two downstream countries act in a compromising manner so that the amount of water they
receive is not reduced.
Figure 8 shows the results associated with the human water-stress criterion. Turkmenistan endures the low-
est tension (i.e., the largest value of this criterion) because it has a smaller population than the other countries.
This means that the amount of water allocated to each person was relatively high, especially that obtained with

Scientific Reports | (2022) 12:3991 | [Link] 8

Vol:.(1234567890)
[Link]/scientificreports/

800.00

700.00

Humean Water Stress (MCM/person)

600.00

500.00

400.00

300.00

200.00

100.00

0.00
Afghanistan Iran Turkmenistan
Country
Initial-Conditions Cooperative Approach Non-Cooperative Approach

Figure 8.  Human water stress criteria for each country (MCM = ­106 ­m3).

0.70

0.60
Agricultural Water Stress

0.50

0.40

0.30

0.20

0.10

0.00
Afghanistan Iran Turkmenistan
Country
Initial-Conditions Cooperative Approach Non-Cooperative Approach

Figure 9.  Agricultural stress criteria for each country.

the initial-conditions water allocation approach. Iran had the smallest value of the human water stress and
this means it endures the highest tension, i.e., the smallest volume of water per capita due to having the largest
population living in the basin.
Figure 9 depicts the results obtained for the agricultural water-stress criterion. Afghanistan features the most
favorable value of this criterion, especially for the non-cooperative approach. In other words, the volume of water
allocated per unit of agricultural land in Afghanistan is the largest.
It is concluded from the data presented in Figs. 7, 8, and 9 that Afghanistan receives the largest agricultural
revenue, the lowest water stresses with respect to human water supply and agricultural water supply under the
non-cooperative approach. It is true that the cooperative approach produces the largest total agricultural revenue
(Fig. 6). At the same time, the benefits of each country demonstrate that all countries receive less benefits under
this approach. It is obvious that available water is constant and if one country takes more water from the river
another country must take less. Therefore, it is fair to state that no approach is better than others because each
country chooses its preferred approach based on its benefits. The actual situation in the Harirud basin is that the
upstream country enjoys the largest water withdrawal, while Iran and Turkmenistan have adopted compromise
strategies in the hope of changing Afghanistan’s strategy.

Conclusion
This study developed a mathematical programming model for water allocation to maximize the interests of the
countries in a transboundary river basin and apply restrictions on their uses of the transboundary river water.
The river system features one upstream country and two downstream countries. The volumes of water available
to meet the countries’ urban and agricultural water demands are considered in this study.
The GA was applied to allocate the transboundary river water. The initial-conditions water allocation was
performed by maximizing water supply reliability. The percentages of the Harirud River’s flow allocated to
Afghanistan, Iran, and Turkmenistan equalled 34, 40, and 26, respectively.
Water allocation was carried out with a cooperative game theory approach to maximize agricultural revenue.
The shares of river flow allocated to Afghanistan, Iran, and Turkmenistan were 36, 42, and 22%, respectively.

Scientific Reports | (2022) 12:3991 | [Link] 9

Vol.:(0123456789)
 [Link]/scientificreports/

Turkmenistan was the most likely country to leave the coalition. Iran’s Shapley value was higher than those of
the other countries in the coalition because its contribution to the coalition’s revenue profits is the largest by
virtue of its large agricultural output.
River water allocation was also calculated with a non-cooperative game theory approach. Countries could
choose to be compromising or oppositional. Afghanistan releases more water downstream when taking the
compromise strategy, and increases the amount of water withdrawn for itself when it chooses to be oppositional.
Iran and Turkmenistan pay incentives to lure Afghanistan to be compromising and deprive Afghanistan of this
benefit if it chooses to be oppositional.
The Nash equilibrium outcome of each country was determined by entering the results obtained from the
combination of the countries’ strategies in the payoff matrix. The Nash equilibrium in this matrix occurs when
Afghanistan is oppositional and the other two countries are compromising. This situation gives rise to a Nash
equilibrium under non-cooperation in which the water allocations to Afghanistan, Iran, and Turkmenistan are
42, 38, and 20% of the river’s flow, respectively.
The allocations were evaluated using stress criteria. The human water-stress criterion indicates that Iran, due
to having the largest population in the Harirud River basin, endures stress. Instead, Turkmenistan endures the
least stress. The agricultural stress criterion was also calculated by the three water-allocation approaches, and
the results showed that the amount of water allocated per unit of agricultural land in Afghanistan is the largest.
Iran and Turkmenistan face relatively similar agricultural stress. The results concerning the evaluation criteria
indicate the best revenues for Turkmenistan, Iran, and Afghanistan are obtained with the initial-conditions water
allocation, the cooperative, and the non-cooperative approaches respectively.
This study’s results demonstrate that transboundary river water must consider all probable strategies and all
possible approaches. In this manner, countries can strategize based on the benefits and costs associated with
probable strategies. The results obtained for the Harirud Basin explain the strategies followed by Afghanistan,
Iran, and Turkmenistan in recent years. Afghanistan earns the most revenues from water use dictated by a
non-cooperative approach. Iran attempts to meet its water needs with a cooperative approach. Turkmenistan
is satisfied with the current situation and cooperates with Iran because it currently receives 50% of the water
downstream (more than the share it would receive if it follows different strategies).
This study showed that the allocation results for water allocation of a transboundary river basin will change
depending on the variable behavior of countries and the uncertainty in the decisions of countries located in the
basin. Results denoted that the upstream country is not willing to cooperate because of the benefits of maximum
consumption. Also, the downstream country, which is larger in terms of agricultural economy and number of
inhabitants in the region and irrigable land area, tends to cooperate with the upstream country in order to meet
the needs of its region. In other words, for further research, it is necessary to add other parameters that are
benefits other than the benefits of water consumption to the issue of allocation so that all countries may become
cooperative. In the current situation, the Harirud basin does not have sustainable management of water resources
and the environment, and probably will face more serious problems in the future for all riparian countries.
Cooperative management of the transboundary river basin would produce the best long-term outcomes for the
transboundary countries. This would require the countries to make some concessions in exchange for achieving
sustainable water management that would produce benefits for all of them indefinitely.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable
request.

Code availability
The codes that support the findings of this study are available from the corresponding author upon reasonable
request.

Received: 1 December 2021; Accepted: 1 March 2022

References
1. Mirzaei-Nodoushan, F., Bozorg-Haddad, O. & Melesse, A. M. Shared water resources management. In Economical, Political, and
Social Issues in Water Resources (eds Mirzaei-Nodoushan, F. et al.) 153–189 (Elsevier, 2021).
2. Wolf, A. T. Sharing Water, Sharing Benefits: Working Towards Effective Transboundary Water Resources Management (UNESCO,
2010).
3. Karimian, H., Li, Q. & Chen, H. F. Assessing urban sustainable development in Isfahan. Appl. Mech. Mater. 253, 244–248 (2013).
4. Bozorg-Haddad, O., Moradi-Jalal, M., Mirmomeni, M., Kholghi, M. K. H. & Mariño, M. A. Optimal cultivation rules in multi-crop
irrigation areas. Irrig. Drain. 58(1), 38–49 (2009).
5. Akbari-Alashti, H., Bozorg-Haddad, O., Fallah-Mehdipour, E. & Mariño, M. A. Multi-reservoir real-time operation rules: A new
genetic programming approach. Proc. Inst. Civ. Eng. Water Manage. 167(10), 561–576 (2014).
6. Kazemi, M., Bozorg-Haddad, O., Fallah-Mehdipour, E. & Loáiciga, H. A. Inter-basin hydropolitics for optimal water resources
allocation. Environ. Monit. Assess. 192(7), 1–10 (2020).
7. Von Neumann, J. & Morgenstern, O. Theory of Games and Economic Behavior (Princeton University Press, 1944).
8. Nash, J. F. Equilibrium points in N-person games. Proc. Natl. Acad. Sci. 36, 48–49 (1950).
9. Holt, C. A. & Roth, A. E. The Nash equilibrium: A perspective. Proc. Natl. Acad. Sci. 101(12), 3999–4002 (2004).
10. Harrison, G. W. & List, J. A. Field experiments. J. Econ. Lit. 42(4), 1009–1055 (2004).
11. Degefu, D. M., He, W., Yuan, L., Min, A. & Zhang, Q. Bankruptcy to surplus: Sharing transboundary river basin’s water under
scarcity. Water Resour. Manage. 32(8), 2735–2751 (2018).
12. Arfanuzzaman, Md. & Abu Syed, M. Water demand and ecosystem nexus in the transboundary river basin: A zero-sum game.
Environ. Dev. Sustain. 20(2), 963–974 (2018).

Scientific Reports | (2022) 12:3991 | [Link] 10

Vol:.(1234567890)
[Link]/scientificreports/

13. Zeng, Y., Li, J., Cai, Y., Tan, Q. & Dai, C. A hybrid game theory and mathematical programming model for solving trans-boundary
water conflicts. J. Hydrol. 570(1), 666–681 (2019).
14. Qin, J., Fu, X., Peng, S. & Huang, S. An integrated decision support framework for incorporating fairness and stability concerns
into river water allocation. Water Resour. Manage. 34(1), 211–230 (2020).
15. Bergström, S. & Göran, L. Interpretation of runoff processes in hydrological modelling-experience from the HBV approach. Hydrol.
Process. 29(16), 3535–3545 (2015).
16. Yang, X., Magnusson, J., Huang, S., Beldring, S. & Xu, C. Y. Dependence of regionalization methods on the complexity of hydro-
logical models in multiple climatic regions. J. Hydrol. 582, 124357 (2020).
17. Seibert, J. & Beven, K. J. Gauging the ungauged basin: How many discharge measurements are needed? Hydrol. Earth Syst. Sci.
13(6), 883–892 (2009).
18. Samuel, J., Coulibaly, P. & Metcalfe, R. A. Estimation of continuous streamflow in Ontario ungauged basins: Comparison of
regionalization methods. J. Hydrol. Eng. 16(5), 447–459 (2011).
19. Pool, S., Viviroli, D. & Seibert, J. Prediction of hydrographs and flow-duration curves in almost ungauged catchments: Which
runoff measurements are most informative for model calibration? J. Hydrol. 554, 613–622 (2017).
20. Holland, J. Adaptation in Natural and Artificial Systems (MIT Press, 1975).
21. Bozorg-Haddad, O., Solgi, R. & Loáiciga, H. A. Evolutionary and Metaheuristic Algorithms in Engineering (Wiley, 2017).
22. Shapley, L. S. A value for n-person games. In Contributions to the Theory of Games II. Annals of Mathematic Studies Vol. 28 (eds
Kuhn, H. W. & Tucker, A. W.) 307–317 (Princeton University Press, 1953).
23. Hooper, B. P. & Lloyd, G. J. Report on IWRM in transboundary basins. In UNEP-DHI Cent. Water Environ (2011).
24. Gately, D. Sharing the gains from regional cooperation: A game theoretic application to planning investment in electric power.
Int. Econ. Rev. 15(1), 195–208 (1974).

Acknowledgements
The authors thank Iran’s National Science Foundation (INSF) for its support for this research.

Author contributions
F.M.-N.; First author; Data curation; Investigation; Formal analysis; Resources; Roles/Writing—original draft.
O.B.-H.; Second author; Corresponding author, Conceptualization; Funding acquisition; Methodology; Project
administration; Supervision; Validation; Visualization; Roles/Writing—original draft. H.A.L.; Forth author;
Validation; Visualization; Writing—review & editing.

Competing interests
The authors declare no competing interests.

Additional information
Correspondence and requests for materials should be addressed to O.B.-H.
Reprints and permissions information is available at [Link]/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit [Link]

© The Author(s) 2022

Scientific Reports | (2022) 12:3991 | [Link] 11

Vol.:(0123456789)