## Introduction

Freshwater and electricity are essential inputs to many production activities that drive the economic development and well-being of societies. The scarcity and variability of freshwater resources have been shown to affect the economic growth of nations1,2. Empirical evidence of uni- and bidirectional relationships between energy consumption and economic development have been documented in countries around the world3,4,5,6. Water and energy systems are interlinked with each other and with several sectors, including agriculture and industry7. Globally, hydropower contributes around 16% of electricity generation and approximately 70% of renewable electricity generation8, and rivers are frequently used to cool power plants9,10. Energy is used for water treatment, pumping, and desalinization. Therefore, efficient use of limited water resources to achieve sustainable economic development requires assessing water and economy interventions in an integrated way.

The Nile is one of the longest rivers in the world and has a basin that extends over 11 African countries, each with a different contribution to and economic dependence on the river11,12. The Nile comprises three main tributaries: the White Nile, the Blue Nile, and the Tekeze-Atbara (see Fig. 1). The Blue Nile originates in Ethiopia and contributes around 57% of the Nile streamflow as measured near the Sudanese–Egyptian border13. High inter- and intra-annual variabilities characterize the Blue Nile streamflow, with around 80% of the flow occurring from July to October14. Nearly all the Nile streamflow, measured near the Sudanese–Egyptian border, is currently consumed by the two most downstream riparian countries, i.e., Egypt and Sudan. Egypt’s water, energy, food, and economic system is linked to the Nile streamflow, which provides around 90% of the country’s freshwater consumption15 and 7% of its electricity supply through hydropower8. On average, irrigated agriculture accounts for approximately 82% of Egypt’s annual Nile water withdrawal, while municipal and industrial water users account for 18%16.

In 1999, the Nile Basin countries established the Nile Basin Initiative (NBI) as a forum for coordination and collaboration on managing the river17. The NBI refers to two regions of the Nile Basin: the Eastern Nile Basin (Fig. 1) and the Nile Equatorial Lakes Region. The NBI worked with the member states to craft the Nile River Basin Cooperative Framework Agreement (CFA)18. In 2010, Egypt and Sudan froze their memberships in the NBI due to disagreements over the text of the CFA, but Sudan returned to full membership 2 years later19.

The United States Bureau of Reclamation conducted a study between 1958 and 1963 (published in 1964) that identified potential dams and irrigation projects on the Blue Nile in Ethiopia20. Dams on the Blue Nile in Ethiopia could increase hydropower generation markedly21. However, they could increase the complexity of managing multiyear droughts20,22. Under the auspices of the NBI, Ethiopia, Sudan, and Egypt launched the Joint Multipurpose Program in 2005 to facilitate coordinated development of investment projects in the Eastern Nile Basin. The three countries invited the World Bank to constitute an independent expert team to conduct a scoping study to inform the selection of a first set of multipurpose projects that would benefit all three countries22. The scoping study team concluded that the best investment opportunities for a joint multipurpose project were in the Blue Nile Basin in Ethiopia. Egyptian policymakers and technical experts contested this conclusion and argued that there were promising options for a joint multipurpose project in the Baro–Akobo–Sobat Basin23. This disagreement on the scoping study’s main conclusion led to a loss of political momentum and eventually failure to achieve the JMP goals23. In 2011, Ethiopia unilaterally started the construction of the Grand Ethiopian Renaissance Dam (GERD). On 12 July 2020, construction was around 75% complete, and Ethiopia began the initial filling of the dam’s reservoir.

When completed, the GERD will be the largest hydropower facility in Africa, with a power capacity of 5150 MW and reservoir storage of 74 billion cubic meters (bcm). The dam will double Ethiopia’s electricity generation and potentially stimulate the country’s economic growth through increases in the output of electricity-dependent sectors and other sectors via forward and backward economic linkages. However, the initial filling and long-term operation of the GERD reservoir are expected to significantly alter the pattern of flow of the Blue Nile downstream of the dam, imposing a range of opportunities and risks to Sudan and Egypt24,25,26. Sudan is expected to benefit from the GERD in terms of improved irrigation water supply reliability, hydropower generation, and riverine flood control provided there is essential daily coordination and data sharing with Ethiopia24,27, but adverse environmental impacts and a loss of recession agriculture are also anticipated28.

Despite several years of negotiations (since 2011) between the Ethiopian, Sudanese, and Egyptian governments on the GERD’s initial filling and long-term operation, no agreement has yet been reached. In 2012, the three countries formed an International Panel of Experts (IPoE) to review the design and impact reports of the dam29. The IPoE recommended conducting further technical studies on the GERD design and impacts. An international, non-partisan Eastern Nile working group met at the Massachusetts Institute of Technology in 2014 to discuss the impacts of the GERD on regional development30. The group pointed out four issues: the need for coordinated operation of the GERD and the High Aswan Dam (HAD), technical concerns on the design of the GERD, the need for an electricity sale agreement, and potential negative impacts on irrigated agriculture in Egypt and recession agriculture in Sudan. In March 2015, the heads of the Ethiopian, Sudanese, and Egyptian governments signed a Declaration of Principles (DoP) on the GERD31. Most of the principles in the DoP are derived from the United Nations Convention on the Law of the Non-navigational Uses of International Watercourses32, including the principles of equitable and reasonable utilization and not to cause significant harm. The DoP stated the need to conduct the studies recommended by the IPoE and utilize the outcomes of these studies to agree on rules and guidelines for the initial filling and long-term operation of the GERD. In 2015, the three countries agreed to contract a consortium of international consultancy firms to conduct the studies recommended by the IPoE. However, this effort failed due to disagreements among the riparians on the terms of reference of the studies and how baseline water allocations should be handled in the construction of scenarios to be examined in the consortium’s analyses33.

In 2018, Ethiopia, Sudan, and Egypt formed a National Independent Scientific Research Group (NISRG) of researchers from the three countries. The NISRG process did not lead to a final agreement among the riparians, but its technical outcomes constituted the basis for subsequent negotiations on the initial filling and long-term operation of the dam34. From November 2019 to February 2020, several rounds of negotiations occurred, including meetings in Washington DC with the United States Government and the World Bank as observers35; yet, no agreement was signed. From June 2020, the African Union initiated and hosted further negotiations between Ethiopia, Sudan, and Egypt with the United States Government, the World Bank, and the European Union as observers. Still, no agreement had been reached at the time this article was being finalized (September 2021). The main points of contention remaining among the riparians are (1) the length of the agreement, i.e., whether it is an interim or permanent agreement; (2) the relationship between the GERD agreement and future water development projects in Ethiopia, (3) the linkage between the GERD agreement and a permanent water allocation agreement among the riparians in the Nile Basin, and (4) the mechanisms to resolve future conflicts should they arise34.

Previous studies have investigated the impacts of GERD filling and long-term operation on Ethiopia, Sudan, and Egypt24,25,26,27,36,37,38,39,40. However, these studies used simple representations of the linkages between the river system and the Eastern Nile economies. In reality, economic growth affects water and electricity demands, and the abundance or scarcity of water and electricity have a feedback effect on economic growth41.

We have two objectives in this paper. The first is to present a new coevolutionary hydro-economic modeling framework that captures the dynamic interactions between a river’s hydrology and infrastructure and the macroeconomy of one of the river’s riparians (Egypt). The second is to use this multi-sector dynamic modeling framework to examine a coordinated operating strategy (termed “coordinated operation”) for filling and operating the GERD on the Nile River. With the coordinated operation, the GERD helps, under specific conditions, meet water consumption targets in Egypt and Sudan, and Ethiopia takes advantage of extra water during periods of high flows to increase GERD storage and maximize hydropower production. Coordinated operation is best conceptualized as an operating policy of “neighbors looking out for each other,” especially during multiyear hydrological droughts. We compare the coordinated operation strategy to a strategy that resembles the proposal for operating the GERD discussed as part of the negotiations between Egypt, Sudan, and Ethiopia in late 2019 and early 2020 in Washington DC (herein, the examined strategy, including some assumptions, is termed the “Washington draft proposal”). Details on the assumptions and differences between coordinated operation and the GERD operating policy in the Washington draft proposal are provided in the next section and the “Methods” section. To assess the dynamic interactions between the Nile and the sectors of Egypt’s economy, we use a calibrated river system model of the Eastern Nile Basin coupled with a Computable General Equilibrium (CGE) model of Egypt’s economy. The water and economy models are developed and connected using open-source modeling frameworks42,43,44, as described in the “Methods” section. Results show that in most of the examined hydrological scenarios, coordinated filling and operation of the GERD increases the total electricity generation from both the GERD and the entire Nile system, sustains Sudan’s Nile water use, decreases Egypt’s irrigation water deficits, and increases Egypt’s total gross domestic product (GDP) and other macroeconomic metrics compared to the Washington draft proposal.

## Results

### The Washington draft proposal versus coordinated operation

We compare the impacts of two GERD filling and long-term operation approaches: (a) Washington draft proposal and (b) coordinated operation. Table 1 summarizes the two examined operating strategies and their key assumptions. Further details on how the two operating strategies are implemented in the modeling framework are provided in the “Methods” section. Both operating strategies assume that Ethiopia targets withdrawal of 2.5 bcm annually upstream of the GERD and that Sudan targets withdrawal of 17.7 bcm/year. The assumed total Ethiopian water withdrawal target is the sum of the withdrawal targets of the Finchaa and Beles irrigation schemes, which are on the Blue Nile, whereas the total Sudanese water withdrawal target is the sum of the withdrawal targets of existing irrigation and municipal water users in Sudan on the Blue Nile, the White Nile, the Tekeze-Atbara, and the Main Nile. Egypt attempts to withdraw 3.8 bcm upstream of the HAD and release 51.7 bcm from the HAD (a total of 55.5 bcm), its water allocation under the 1959 Nile Waters Agreement45. Deficits in Egypt are measured from this 55.5 bcm target. It is worth noting that Egypt and Sudan have different views on how evaporation losses should be considered in their 1959 bi-lateral water allocation agreement45. Egypt believes that reservoir evaporation from dams constructed after the 1959 agreement (i.e., Merowe, Roseires heightening, and Upper Atbara and Setit) is part of water allocations, while Sudan argues the opposite25. In this study, we assume the target withdrawals by Ethiopia and Sudan and the target releases from the HAD to illustrate the behavior of the hydrological and economic systems; they do not reflect an endorsement of the status quo water allocation in the Nile Basin.

In this study, the examined Washington draft proposal resembles the proposal annexed in the letter of the permanent representative of Egypt to the United Nations to the President of the United Nations Security Council dated 19 June 202046. Ethiopia has not accepted this draft proposal. The Washington draft proposal suggests limiting the period during which Ethiopia can retain Nile flows to fill the GERD Reservoir to the peak of the annual flood season in July and August35. The Washington draft proposal enables Ethiopia to ramp up GERD storage so that all turbines can become operational within the first 2 years of initial filling to guarantee that Ethiopia can quickly begin electricity generation. The Washington draft proposal mitigates the consequences of droughts, prolonged droughts, and prolonged periods of dry years using three operating constraints on the GERD46: (a) a minimum annual release depending on the inflow, (b) a 4-year minimum average annual release, and (c) a 5-year minimum average annual release. Given these three constraints on GERD operation, Ethiopia would still have some operational flexibility. Thus, we assume that once reservoir storage reaches the long-term operation stage (49.3 bcm), Ethiopia would operate the GERD to maximize the 90% power reliability and sustain a minimum environmental flow of 43 Mm3/day (million cubic meters/day) subject to these three drought mitigation mechanisms. Further details on the drought mitigation mechanisms of the Washington draft proposal and their implementation and assumptions in the model are provided in the “Methods” section.

Figure 2 shows a flowchart of the examined coordinated operation strategy. A description and a schematic of the Eastern Nile River system model are provided in the “Methods” section and the Supplementary information. The rationale behind the coordinated operation strategy we examine in this study aims for some measure of hydro-solidarity, “neighbors looking out for each other,” where political boundaries are relaxed but with some national goals remaining. The coordinated operation strategy is not designed to be a prescribed solution for Nile water issues; rather, we use it to demonstrate the direction of change in economic performance and resilience that transboundary collaboration holds. We configured the coordinated operation strategy manually through trial and error, inspired by operating policies of previous studies24,37,47. In the coordinated operation approach, each of the three countries has a role to play. Ethiopia has more flexibility in GERD management when there is sufficient water in the HAD reservoir (HADR) and when the flows of the other two Eastern Nile tributaries (i.e., the White Nile and Tekeze-Atabra; Fig. 1) are high. In addition, Egypt shares information with Ethiopia on HAD storage and target water release. Coordinated operation enables Ethiopia to avoid the constraints on minimum releases from the GERD that are part of the Washington draft proposal. Instead, water releases from the GERD always ensure the satisfaction of water consumption targets on the Blue Nile and Main Nile in Sudan, and when physically possible and HAD storage is low (below 50 bcm), the satisfaction of water consumption targets in Egypt, with Ethiopia able to seek additional benefits under favorable streamflow conditions. A HADR storage of 50 bcm is equivalent to a reservoir water level of 156 m a.s.l., which is 9 m above the turbine shutdown level of the dam24. With thecoordinated operation, the operations of the Roseires, Sennar, and Merowe dams located in Sudan between the GERD and Egypt have been adapted to pass water releases from the GERD intended to reach Egypt. Ethiopia is assumed to operate the GERD to maximize the 90% power reliability and sustain a minimum environmental flow of 43 Mm3/day, subject to constraints that water consumption targets in Sudan and Egypt are satisfied under specific conditions.

These two filling and long-term operation approaches are analyzed using 102 different 30-year river flow sequences (traces) developed using the index-sequential method48. This method generates river flow traces from the historical flow record, taking every year in the record as a possible starting point. We used the 1901–2002 Nile flow data to generate the river flow traces (see Supplementary Fig. 1).

Sudan receives irrigation, flood control, and hydropower benefits from the GERD, assuming daily coordination and active data sharing between the GERD and Roseires Dam. The modeling results show that these benefits are essentially the same with the Washington draft proposal and the coordinated operation. This is due to Sudan’s geographic advantage of being located upstream of Egypt and the relatively small storage dams and hydropower capacity in the country. Furthermore, we assume that the adverse environmental impacts and the loss of recession agriculture in Sudan are similar in the two examined GERD operation scenarios. Therefore, we only present the results of the impacts of the coordinated operation for Ethiopia and Egypt.

### Coordination can improve water utilization

Figure 3 illustrates the change in Nile water withdrawal in Egypt, hydropower generation of the GERD and Egypt, and the total reservoir evaporation, Toshka spills, and river channel seepage as a result of coordinated operation compared to the Washington draft proposal. Table 2 reports statistics for some of the metrics shown in Fig. 3. Results show that in 77% of the traces simulated,  the coordinated operation would decrease Egypt’s total water deficits compared to the Washington draft proposal. Most of the significant decreases in irrigation water deficits occur after 2025 because the HADR is currently full49 and can satisfy any near-term Egyptian water supply deficits that may occur in a specific simulation. Supplementary Fig. 2 shows that the decreases in Egypt’s irrigation deficits occur during multiyear periods of water scarcity. Supplementary Fig. 2 also shows a drawdown of GERD storage with the coordinated operation to help alleviate irrigation water deficits in Egypt when HAD storage falls below 50 bcm. This decline in GERD storage resulted in a small reduction in the dam’s total energy generation.

Over 2020–2049, 58% of modeled traces show a decrease in the total Egyptian hydropower generation in the coordinated operating strategy compared to the Washington draft proposal. Most of the significant annual declines in Egypt’s hydropower happen from 2020 to 2030 due to the faster decrease in HADR storage with the coordinated operation that results from rapid GERD initial filling, which is enabled by the HADR having enough initial storage at the start of the simulation to meet Egypt’s target releases. In some simulated traces, electricity generation from the HAD also declines beyond 2030 with the coordinated operation compared to the Washington draft proposal. This occurs because, in some simulated traces, the coordinated operation shifts water storage from the HADR to GERD Reservoir, resulting in lower water levels at the HAD and higher water levels at the GERD.

Figures 3d, e show the effect of the coordinated operation on GERD and system-wide hydropower generation, respectively, compared to the Washington draft proposal. The time-series plots show some declines in GERD electricity within the first 5 years, followed by increases. The declines occur because of lower downstream water releases in the first 5 years to speed up reservoir filling when the storage in the HADR is sufficient to supplement river flows and thus avoid irrigation deficits in Egypt. The long-term increase in GERD electricity generation (e.g., the cyan-colored line in Fig. 3d) results from the faster initial filling and the opportunistic long-term management of dam releases in coordination with Egypt and Sudan. Supplementary Fig. 3 shows how coordinated operation can result in higher reservoir storage at GERD without adversely affecting Egypt’s water consumers. Seventy-one percent of the simulated traces show an increase in the GERD’s cumulative electricity generation over 2020–2049 compared to the Washington draft proposal. The change in the GERD’s total electricity generation over the 2020–2049 simulation period ranges from −1900 to 17,900 GWh, with a median of around 1600 GWh. The changes in the GERD’s annual electricity range from −35 to 84% compared to the operating policy of the Washington draft proposal, with a median annual change of 0%, increases in 40% of the years, and decreases in 10% of the years. Figure 3f shows the impact of the coordinated operation on total water losses from reservoir evaporation, spills to the Toshka depression in Egypt, and channel seepage compared to the Washington draft proposal. The coordinated operation changed cumulative water losses over the 30-year simulation by −18.1 to 1.1 bcm depending on the hydrological trace, with a median of around −5.1 bcm.

### Coordination can enhance economic resilience

Nile flows to Egypt play a vital role in the country’s economy. The agriculture sector accounts for around 23% of the country’s employment50. Changes to irrigation water supply affect the output of agriculture and the livelihoods and employment of millions of Egyptians. Moreover, changes to irrigation water availability affect other economic activities that use agricultural products as intermediate inputs due to forward and backward economic linkages. Although hydropower contributes only around 7% to the Egyptian electricity mix, reduction in hydropower generation increases the use of other electricity generation technologies with higher variable costs. Figure 4a–e depicts the change in some macroeconomic metrics of the Egyptian economy as a result of  the coordinated operation compared to the Washington draft proposal. Table 2 reports key statistics of some of these metrics from the model simulations. Overall, the Egyptian GDP, investment, exports, imports, and government savings would increase as a result of the coordinated operation. The first 5 years of all model simulations show an insignificant change in macroeconomic performance due to the coordinated operation because the HADR starts the simulation full and can supplement reduced inflows to the HADR due to GERD filling and thus Egypt’s water needs can continue to be met. The slight decline in macroeconomic performance during the filling period is due to a reduction in Egypt’s hydropower generation as a result of a speedup of GERD filling under coordinated operation compared to the Washington draft proposal (Fig. 3c). The positive changes in Egyptian macroeconomic performance beyond 2030 are due to improved irrigation water supply.

As the economy-wide modeling results show, the direct and indirect impacts of reduced irrigation deficits increase the production of all economic activities and increase investment (Fig. 4b). Investment grows due to an increase in the savings of households, enterprises, and the Government of Egypt. Imports and exports also increase with coordinated filling and long-term operation of the GERD (Figs. 4c, d). Results for the coordinated operation show an increase in government savings compared to the Washington draft proposal (Fig. 4e). Government income increases as a result of the indirect impacts of improvements in irrigation water supply. Fewer and smaller irrigation water supply deficits lead to an increase in the production of many industries, which increases tax revenues. Moreover, the increase in industrial production raises household income and demand for commodities, imports, and import duties. Government savings increase with coordinated operation as a result of the increase in government income. With the coordinated operation, government spending increases; the increase in government spending is lower than the increase in government income resulting in a net positive increase in government savings. Results show that the change in the present value of Egypt’s GDP over the 2020–2049 simulated traces ranges between US$−0.7 and US$ 4.1 billion with a median of around US$0.24 billion (at a 3% discount rate) and an increase in about 76% of the examined hydrologic scenarios. Egypt’s investment, exports, imports, and government savings follow a similar pattern to that of GDP with median present value changes of around US$ 80, 70, 70, and 20 million, respectively, if the coordinated operation is adopted. Most of the improvements in Egypt’s macroeconomic performance occur during multiyear periods of water scarcity (5–15 years of the 30-year simulation period) when HAD storage is below 50 bcm. The low median change in economic performance indicates that in a large proportion of the simulated traces, the HADR does not drop to a level that triggers GERD’s help.