Should Egypt be afraid of the Grand Ethiopian Renaissance Dam? The consequences of adversarial water policy on the Blue Nile

The development of large dams on a river can serve many purposes, such as the regulation of flows for enhancing agricultural or municipal supplies, reducing the risk of downstream flooding, generating hydropower, creating fishery and navigation opportunities, or modifying ecological systems. The priority placed on different objectives is not only a technical or economic decision but has important political implications for the riparian countries. On a transboundary river, riparians' interests often diverge and can present both a possibility for conflict as well as an opportunity for cooperation (Zeitoun & Mirumachi, 2008). In some circumstances, the management of water can be ‘weaponized’ with one actor intentionally harming another by holding back flows when water is needed, or by making large releases that intentionally cause flooding (Gleick, 2019). Such deliberate adversarial water management actions are historically rare, and distinctly different from the more frequent circumstance when infrastructure is targeted during a military conflict (Gleick & Shimabuku, 2023). The risk of adversarial water management exists in the Nile Basin, but whilst the risk is perceived to be increasing, it has not yet been rigorously characterized.

In 2011, Meles Zenawi, the late Prime Minister of Ethiopia, announced Ethiopia's intention to build the Grand Ethiopian Renaissance Dam (GERD) on the Blue Nile 15 km upstream of the Ethiopian-Sudanese border. Since then, Egypt, Ethiopia, and Sudan have been engaged in sporadic negotiations about the filling and operation of the GERD, without reaching an agreement. Ethiopia began unilaterally filling the GERD Reservoir in 2020. By the summer of 2024, Ethiopia was able to retain 49.3 billion cubic meters (a pool elevation 625 masl) to complete the initial filling process. Although the maximum capacity of the GERD Reservoir is 74 billion cubic meters, any water stored above 49.3 billion cubic meters is expected to be discharged annually by June 30th to create space to manage the flood of the subsequent water year. Despite significant downstream concerns, this filling process did not cause appreciable harm to Egypt. The High Aswan Dam (HAD) Reservoir remains full due to prudent planning efforts by the Egyptian government and high annual flows during the filling period (see Supplementary Materials). Although Sudan faced challenges to manage its own water supply infrastructure during the initial years of the filling process, improved forecasting methods and technical communications with Ethiopia allowed Sudanese water managers to adapt in later years (Hassan et al., 2024).

From time immemorial, Egypt has feared the prospect of Ethiopia holding back the Nile waters and has perceived this possibility as an existential risk (Erlich, 2002). Since the construction of the GERD, this fear has become widespread in Egyptian civil society (Seide & Fantini, 2023), often inflamed by biased media coverage (Elsoufy, 2024) and fed by misleading science (Wheeler & Hussein, 2021; Wheeler et al., 2022; Waktola, 2024). However, there are no analyses in the published literature that quantify how much harm Ethiopia could inflict on Egypt if it decided to operate the GERD for this purpose. Reaching an agreement on how the GERD could be operated to best balance the interests of all three countries requires a realistic assessment of the risks to Egypt associated with adversarial behavior by Ethiopia and an understanding of the benefits of a compromise operating strategy.

In this paper, we examine three stylized approaches that Ethiopia could adopt to operate the GERD. The first policy (labeled ‘Self-Interest’) would be for the GERD to be operated solely to maximize hydropower production, without regard for the consequences (positive or negative) for the downstream riparians.

The second policy (labeled ‘Compromise’) would be to operate the GERD to balance Ethiopia's interests with those of downstream riparian countries. Such cooperative behavior can be conceptualized in several ways. Some scholars imagine cooperation in terms of basin-wide systems optimization in which total economic benefits are maximized and subsequently shared, sometimes described as a ‘Share Benefits, Not Water’ strategy (Sadoff & Grey, 2002; Sadoff & Grey, 2005; Wolf, 2010; Arjoon et al., 2016). However, such concepts are difficult, if not impossible, to operationalize, particularly in a highly securitized and water-scarce context (Phillips et al., 2006). Therefore, in this paper we take a pragmatic approach, conceptualizing a Compromise that entails operating the GERD to assist downstream riparians during extreme multi-year droughts by making supplemental water releases from the GERD.

The third policy (labeled ‘Adversarial’) would be to operate the GERD to intentionally inflict harm on downstream riparians (or to maximize the potential to inflict harm). In this study, we focus on the situation of additional water being retained at the GERD during a drought to negatively affect Egypt. Other scenarios could be analyzed that inflict harm on Sudan, such as a sudden discharge of water that would risk the integrity of the Roseries and Sennar dams or withholding water when it is most critical for Sudanese agriculture. However, since our focus is on the concerns of Egypt, we leave these scenarios for future analyses.

The feasibility and relative impacts of these three stylized policies on the riparians' objectives is an empirical question that we examine in this paper. We recognize that multiple objectives may exist concurrently, and objectives may change over time. However, comparing these three GERD operating policies, alongside a counterfactual baseline in which the GERD had not been built, provides insight into the consequences of strategic behavior. We note that each of these policies is based upon assumptions about: (i) the benefits to Ethiopia of hydropower production; (ii) when additional water released from the GERD might be most beneficial in mitigating drought in Egypt; (iii) what aspects of water shortages might be most harmful in Egypt; and (iv) Sudan following its obligations in the 1959 Nile Waters Agreement and not acting to exacerbate or mitigate effects on Egypt (Whittington, 2022; Whittington 2024). The exact quantitative results are to some extent sensitive to the specification of these assumptions, but the qualitative insights are robust.

We use the well-established, widely-cited Eastern Nile RiverWare Model (ENRM) to simulate the counterfactual and these three different operating policies for the GERD (Wheeler et al., 2016, 2018; Kamel et al., 2019; Wheeler et al., 2020; Siddig et al., 2021; Wheeler et al., 2022; Murgatroyd et al., 2024). This model represents the Blue Nile, White Nile downstream of the Sudd wetlands, the Tekeze-Setit-Atbara Sub-basin and the Baro-Akobo-Sobat Sub-basin. The logic in the model simulates the available and assumed operational policies of the major water management infrastructure in the Eastern Nile Basin including Lake Tana and the Tana-Beles hydropower project, Finchaa Dam, Tekeze Dam, GERD, Roseries Dam, Sennar Dam, Merowe Dam, Jebel Aulia, Khashm El Girba, the dual Upper Atbara and Setit complex, and the HAD (Wheeler et al., 2016, 2018). Following the completion of the GERD and the reduced sediment immediately downstream of the GERD, the operation of the Sudanese Roseries and Sennar reservoirs is assumed to make releases to meet Sudanese demands but otherwise operate fully to minimize the possibility of shortages. The operation of the Merowe Dam is assumed to continue operating primarily for hydropower purposes and seasonal flood management practices.

To understand the implications of a wide range of possible hydrological conditions, we simulate the behavior of the Nile system using 100 different realizations of a 49-year synthetic hydrological flow series. These flow series have been generated from a multi-site stochastic streamflow generator that has been optimized to reproduce the relevant observed flow statistics, including the frequency and duration statistics of low flows (Borgomeo et al., 2015; Wheeler et al., 2018, 2025). The attraction of this approach is that it enables the exploration of a full range of hydrological conditions, including flow sequences and extremes that have not been observed but are consistent with observed statistics.

We show the consequences of the three operating policies in terms of several performance metrics compared to a counterfactual in which the GERD was not built. This comparison illustrates the frequency and severity of water shortages that Egypt would experience even if the GERD had not been built. Using the hundred 49-year synthetic flow series, we first show the pool elevations of both reservoirs when each of the three GERD operating policies is in effect. We then calculate cumulative exceedance probabilities of different annual shortages in Egypt (i.e. probability of an annual shortage being greater than a given value). We report the statistics (for all 100 replicates and 49 years of simulations) of the shortage corresponding to an 0.02 exceedance probability and the worst-case single-year consequences for Egypt. The 0.02 probability corresponds to the likely occurrence of one event in the next 50 years, which approximately matches the 49-year modeling period. The performance indicators for impacts on Egypt (maximum annual and cumulative shortage) are measured relative to a target annual release of 55.5 bcm. These performance measures reflect Egypt's practice to release water according to its stated HAD operating policy that is based on its 55.5 bcm/year claim to Nile water. This does not imply any endorsement of this contested claim by the authors.

For Ethiopia, we show the consequences of the three operating policies in terms of monthly hydropower generation. Cumulative exceedance probabilities estimated over the flow series show the frequency that Ethiopia is able to achieve steady power production.

To graphically illustrate the consequences of the three GERD operating policies, we also use a single historical hydrological flow series (1964–1995) which includes the drought of the 1980s. These simulation results show how the storage of the GERD and HAD Reservoirs, the magnitude and timing of annual water shortages in Egypt, and the annual energy generation of the GERD would vary were this historic flow sequence to be imposed on the three assumed GERD operating policies.

Our simulations assume the initial GERD Reservoir filling is completed and thus begins with an elevation of 638 masl (the expected January elevation during normal operation). In normal operating conditions, the elevation of the GERD Reservoir is allowed to fall in the subsequent months, so that it is below 625 masl at the beginning of each water year (1st of July) for flood control purposes. The management of all the other reservoirs on the Nile reflects our understanding of their current operations (Wheeler et al., 2016, 2018, 2020), including the drought management plan for the HAD that reduces outflows to Egypt by 5, 10, and 15% as storage falls below 60, 55, and 50 bcm, respectively. We assume that existing levels of water withdrawals in Sudan (16.7 bcm/year) and Ethiopia (0.45 bcm/year) are unchanged from current estimates and Egypt seeks to release 55.5 bcm/yearr from the HAD based on their perceived historic rights as specified in 1959 Nile Waters Agreement between Egypt and Sudan.

Previous analysis has shown that 1600 MW of hydropower generation target can be achieved at the GERD with a 90% reliability (Wheeler et al., 2018). This value sets the initial power target for all three operating policies under most hydrologic conditions. We focus on the divergence required from this GERD operational target to meet different objectives during drought conditions.

Ethiopia could operate the GERD in its own financial self-interest by focusing only on maximizing hydropower generation, without regard for any adverse or beneficial consequences to downstream riparians. Nonetheless, even a purely self-interest objective requires a three-way trade-off between (1) maximizing a steady (firm) hydropower generation to use domestically or sell internationally, (2) achieving a specific level of reliability to meet the desired hydropower generation objective, and (3) maximizing the total annual hydropower generation. Raising the hydropower target above 1600 MW to increase domestic or international power sales would require greater releases through the turbines and cause the storage in the GERD Reservoir to decline during subsequent low inflow years. This would risk the reservoir reaching its minimum operation elevation, resulting in generation shortfalls. On the other hand, reducing regular hydropower production below 1600 MW presents a lost opportunity for hydropower revenue but provides a greater reliability. The third objective is to maximize the annual energy generation per cubic meter discharged through the turbines, which would require maintaining the GERD Reservoir nearly full to keep a high head on the turbines. However, keeping a high water level increases the possibility of discharging water over the spillway from which no hydropower is generated. We do not know how Ethiopia will navigate this trade-off between hydropower production, reliability, and production per cubic meter of discharge. From a financial perspective, much will depend on the specifics of future power purchase agreements with neighboring countries and Ethiopia's domestic power requirements.

For the purpose of modeling a Self-Interest operating policy for the GERD, we assume that Ethiopia will maintain the GERD hydropower generation target of 1600 MW during above-average inflow conditions. To avoid generation shortfalls during below-average flows in the Blue Nile, we assume that hydropower generation at the GERD would be reduced incrementally toward 1,000 MW. This would ensure steady, albeit reduced, hydropower generation and allow the GERD storage to remain near its optimum generation elevation but would reduce water releases downstream. Only if an extreme multi-year drought were to occur would the elevation of the GERD Reservoir decline. We assume that the GERD elevation would not be allowed to fall below 610 masl (31 bcm), the minimum desired head of 103 m for the turbines (Edrees, 2020). We show details of this Self-Interest operating policy alongside alternative hydropower and reliability targets in the Supplementary Materials.

An operating policy that Ethiopia could use to foster goodwill and build positive economic and diplomatic relationships would be to release significant quantities of water from GERD storage to help Egypt and Sudan during drought conditions. Such a strategy would increase the reliability of hydropower generation throughout the Nile system. Ethiopia would agree to make supplemental releases from GERD storage to assist Egypt when multi-year inflows to the GERD were far below normal. Ethiopia would still secure a steady revenue stream from hydropower generation in almost all circumstances. Given the GERD's large hydropower generation capacity (5150 MW), we assume Ethiopia can benefit from such releases up to this power generation limit. However, if releases exceed this capacity, additional water would have to be discharged through the two bottom outlets.

Possible versions of cooperative operating policies at the GERD have been extensively studied (Arjoon et al., 2016; Jeuland et al., 2017; Wheeler et al., 2018; Basheer et al., 2021, 2023). The version we adopt here does not represent an idealistic cooperative basin-wide optimum but is a practical compromise that to some extent safeguards downstream interests. To codify the Compromise operating policy in the ENRM, we first assume the GERD operates to meet the 1600 MW hydropower generation objective despite the onset of a drought. If a multi-year drought occurs, additional downstream releases are made through a policy that has two parts:

First, if the average inflows to the GERD Reservoir over a 3-year period fall below 39 bcm/year, Ethiopia would draw down storage in the GERD Reservoir in three annual stages to provide Sudan and Egypt with supplemental water supplies to reduce the consequences of drought conditions. However, storage in the GERD would not be reduced below an elevation of 607 meters above sea level (masl), which is a compromise between Egypt's stated position of 603 masl and Ethiopia's stated position of 610 masl (Edrees, 2020).

Second, after such a prolonged drought the storage in both the HAD Reservoir and the GERD Reservoir would be low, and both reservoirs would need to be refilled. Ethiopia would prefer that the GERD Reservoir be refilled first to increase the head on the turbines and increase turbine efficiency. From a systems perspective, storing water upstream also has the advantage of reducing system-wide evaporation losses (Lund & Guzman, 1999). On the other hand, Egypt would prefer to refill the HAD Reservoir first, with at least enough water to restore hydropower generation at the HAD and ensure available water for immediate and critical uses.

For this Compromise operating policy, we assume that a balanced drought recovery strategy is invoked to share annual inflows when they increase. Ethiopia would not begin refilling the GERD Reservoir as long as forecasted inflows to the GERD Reservoir remained below 37 bcm/year. As the annual inflows increase, we assume the GERD Reservoir would retain half of any water in excess of 37 bcm and the remainder would be released downstream for Sudan and the HAD Reservoir. More specific details of this Compromise operating policy are provided in the Supplementary Materials.

Ethiopia could decide to use the GERD to inflict harm on downstream riparians. It is understandable for the downstream riparians to be concerned about this possibility and to speculate as to whether and how Ethiopia might pursue an Adversarial operating policy.

For Ethiopia to inflict harm on Egypt, two conditions would have to coincide. First, the storage in the HAD Reservoir would need to be low, i.e., near or below the volumes at which Egypt would invoke sequential stages of its drought management policy (DMP). This policy restricts releases from the HAD and thus entails some reduction in Egyptian water use. Second, the GERD Reservoir would have to be able to retain additional water, i.e., to have sufficient empty storage to reduce the total amount of water discharged over a year. The quantity of water that can be withheld would depend on the amount of unused storage, so would require the GERD Reservoir to be below the annual minimum flood space elevation of 625 masl. This storage space would need to be created by prior water releases. Ethiopia could hypothetically maintain the GERD Reservoir at a low level to maximize the potential to withhold water should it wish to inflict harm in Egypt. However, this strategy would conflict with the GERD's capability to efficiently generate steady hydropower.

In the Adversarial operating policy that was implemented in the ENRM, we first assume the GERD operates to meet the 1600 MW hydropower generation objective despite the onset of a drought. We then assume that releases from the GERD would be abruptly ceased when HAD Reservoir storage fell to 70 bcm, which is immediately above the level at which the HAD DMP would be invoked. We recognize that such an operation would have negative impacts in Sudan. However, the focus of our current analysis is on adversarial policies toward Egypt. Due to the GERD generating 1600 MW prior to this point, the GERD storage would inevitably be below 625 masl by the time the storage in the HAD Reservoir reaches 70 bcm, providing Ethiopia with the possibility to abruptly retain water. The Supplementary Materials explain how this HAD Reservoir threshold was selected and the consequences of different thresholds.

Although each of the three operation policies of the GERD result in average evaporation between 1.6 and 1.9 bcm, these losses do not directly map onto shortages in Egypt. The implications for shortages to Egypt are driven primarily by how the various operations of the GERD increase the frequency that the HAD Reservoir falls to 60 bcm, thereby invoking their DMP. In 93.5% of the years modelled, there are no reductions to Egypt's release of 55.5 bcm, even if Ethiopia invoked an Adversarial operating policy.

The difference between the three operating policies (and also the No GERD counterfactual) tend to converge near the 8 bcm level, when the third level of the HAD DMP is invoked (15% × 55.5 bcm = 8.3 bcm). There is a very small number of years in the stochastic flow simulation in which the Adversarial policy is able to inflict more severe shortages, reaching a maximum of 16.6 bcm, relative to 8.3 bcm in the Compromise and Self-Interest operating policies and 8.0 bcm in the No GERD counterfactual. However, these results should be treated with caution, as they are rare in the stochastic flow series – occurring with an annual probability of less than 0.001. However, as described below, this magnitude of shortage is similar to what would result from a reoccurrence of the 1980s drought and thus warrants consideration.

By contrast, because the Adversarial policy involves aggressively filling the reservoir in some conditions, curtailing releases, and risking spills, it provides the least reliable hydropower production. Considering all 100 hydrologic scenarios and 49 years of simulation, the average annual electricity generation from the GERD is 13.28 TWh under the Adversarial operating policy, less than under both the Compromise operating policy (13.48 TWh) and the Self-Interest operating policy (14.26 TWh). Thus, the Adversarial operating policy results in a reduction in GERD hydropower generation of 0.19 TWh per year (1%) compared to the Compromise operating policy and 0.98 TWh per year (7%) compared to the Self-Interest operating policy.

If a kilowatt-hour were valued at US$0.05 and there were always available buyers of electricity generated up to 5150 MW in all scenarios, the cost of the Compromise operating policy relative to the Self-Interest policy would be a financial loss of $39 million per year on average. Furthermore, on average, the cost of the Adversarial policy would be a financial loss of $49 million per year loss relative to the Self-Interest policy. These losses represent the average impact across all simulated years resulting from different policies; however, the differences are significantly larger when assessing losses over a single event such as the drought of the 1980s.

Comparing the Adversarial and Compromise policies, we see that the shortages to Egypt are effectively compacted into a shorter timeframe by Ethiopia operating the GERD in an adversarial way, but at the risk of losing hydropower generation by needing to spill water once the GERD Reservoir is full. Such an adversarial action would result in 28 GWh of lost hydropower generation over an 11-year period, severely impacting Ethiopia's own power production and income from foreign hydropower sales (Figure 4(c)). Using a value of US$0.05 per kilowatt-hour, this adversarial act would cost Ethiopia $1.4 billion in lost hydropower revenues, as shown in the shaded areas of Figure 4(c).

These results show that it is possible to operate the GERD to exacerbate water shortages downstream in Egypt by withholding water during a multi-year drought or by quickly refilling the GERD after a prolonged drought. However, there would be few occasions when Ethiopia would be able to cause harm to Egypt. When the HAD Reservoir is within its normal operating range, it has sufficient water stored to buffer shortfalls in inflow which, if they are caused by GERD operation, are inevitably temporary. The frequency that the HAD Reservoir would be depleted near the 60 bcm storage level (159.4 masl) –which triggers their DMP – is low (less than 6.5%).

For Ethiopia, there are three limitations to implementing an Adversarial operating policy. First, Ethiopia cannot implement this policy anytime it desires. It may take a long time for the opportunity to arise when the storage in the HAD Reservoir and the GERD Reservoir are both low. This opportunity to inflict harm may not coincide with a period of geopolitical tension when Ethiopia would want to inflict harm.

Second, maintaining empty storage in the GERD Reservoir that could be used to withhold water would be costly. From a financial perspective, Ethiopia will generally want to keep storage in the GERD Reservoir high to maximize hydropower generation and reliability, thus not having the additional storage capacity that would be required to capture additional water. Unless Ethiopia intentionally operated the GERD with less storage than optimal, Ethiopia will have little opportunity to hold back significant quantities of water. Ethiopia would pay an immediate price both in terms of reduced hydropower and internal political pressures of not managing the GERD Reservoir within its design operation range.

Third, because Ethiopia cannot forecast future annual inflows to the GERD, it might act to harm Egypt without succeeding. A high inflow to the GERD might arise in the year after Ethiopia acted to withhold water, negating the harm it was trying to inflict.

The only opportunity that Ethiopia has to inflict substantial harm by withholding water occurs during a multi-year drought when the HAD Reservoir has little storage remaining and the GERD Reservoir also has reduced storage. Multi-year droughts are rare, but we find that at the 0.02 annual exceedance level, the Adversarial policy could result in a maximum annual deficit of 5.77 bcm, which is 1.86 bcm worse than the Self-interest policy, 2.72 bcm worse than the Compromise policy, and 4.8 bcm worse than the No GERD counterfactual situation.

Because the GERD Reservoir is a hydropower reservoir with relatively small and steady evaporation losses (1.6 to 1.9 bcm/year, depending on the operating policy), its primary effect is to redistribute downstream flows across time on a sub-annual basis, and to a lesser degree, on a multi-year basis. Thus, the average reduction in downstream releases across decades is equal to the evaporation losses from the GERD Reservoir. However, after the effects of the downstream reservoirs including the buffering effect of the HAD, the GERD adds less than 0.2 bcm of average annual shortages to Egypt. The main factor among the operational policies that result in different impacts to Egypt is how frequently each policy would invoke different levels of the HAD DMP. Only by nefariously operating the GERD concurrently with a multi-year drought can Ethiopia use the GERD to create downstream conditions that are significantly worse than in the counterfactual situation. Even then, the negative consequences would only occur for a few years because the GERD Reservoir would soon thereafter be full and again releasing water downstream to Sudan and Egypt through turbines or over spillways.

If Ethiopia operated the GERD to pursue its financial self-interest (releasing water to meet its hydropower generation objective without concern for the downstream consequences), the risks to Egypt are not very different from GERD operating policies that include some attempt by Ethiopia to release water at times of shortage and recover it after the drought. The Self-Interest operating policy results in the water elevation at the GERD being higher, on average, in which case Ethiopia has limited opportunity to strategically retain water. High water levels at the GERD should assure Egypt that Ethiopia was not pursuing an adversarial operating policy. However, during a multi-year drought, if Ethiopia kept the GERD Reservoir full and did not make supplemental releases to assist the downstream riparians, this could also be perceived negatively in Egypt. Paradoxically, only by reducing storage in the GERD Reservoir to ostensibly support Egypt, could Ethiopia create the opportunity to inflict harm upon Egypt. Our results show that the main benefits of cooperative operating strategies are to ensure against Ethiopian adversarial behavior during a refilling process, not against self-interest behavior.

The riparians should continue their efforts to reach a fair agreement on the GERD operating policy that balances the interests of the parties. A key benefit of an agreement on GERD operating policy would be to reduce the perceived risk of strategic behavior by Ethiopia. A cooperative agreement should not only include supplemental releases from the GERD to assist downstream riparians during a prolonged drought but should have clear guidelines on when and how the GERD and HAD Reservoirs should be refilled. Such guidelines would provide assurance that Ethiopia was not, in fact, acting to deliberately inflict harm. Reaching a binding agreement on this infrequent yet critical issue would greatly reduce the risk of the spread of misinformation and a resulting ‘water panic.’ Our results show that the costs to Ethiopia of such a cooperative agreement in terms of foregone hydropower generation would be minimal, particularly if all water released could be used to generate hydropower.

The most substantial increased risk to Egypt of future water deficits arises not from decisions about how the GERD will be operated, but rather from the magnitude of increased upstream withdrawals in Sudan and Ethiopia (Murgatroyd et al., 2024). Our analysis has demonstrated that Egyptian fears about the operation of the GERD for hydropower production are largely misplaced. However, there could be a risk of riparians misreading each other's intentions in a multi-year drought. It is thus important to analyze the possibility of Ethiopia using an adversarial GERD operating policy to explicitly address this understandable fear within Egyptian civil society and policymakers. When misinformation is easily spread by social media, unfounded fears can lead to a ‘water panic’ (Wheeler et al., 2020; Wheeler & Hussein, 2021). By studying the conditions under which Ethiopia could operate the GERD to inflict harm, the riparians can more effectively avoid any misunderstanding about each other's intentions. Understanding when it is possible for Ethiopia to inflict harm should enable Ethiopia to clearly signal to downstream riparians that it is not operating the GERD in this manner, thus reducing the risk of a ‘water panic.’

D.W., J.H, A.M, and K.W. contributed to the formulation of the concepts for this study. K.W. and A.M. contributed the modeling analysis. D.W., J.H, A.M., and K.W. contributed to writing and editing the manuscript.

Materials and Correspondence should be directed to Dr Kevin Wheeler

Dale Whittington was a member of the Eastern Nile Scoping Study Team from 2006 to 2009, whose work was commissioned by the Eastern Nile Council of Ministers and funded by the World Bank. In 2014–2015, he participated as a member of the International, Non-partisan Eastern Nile Working Group, convened by the Massachusetts Institute of Technology. In 2017–2018 he served as a member of an external technical advisory committee for Deltares on a project – financed by the Nile Water Sector, Government of Egypt – to study the effects of the Grand Ethiopian Renaissance Dam on the Nile system. He served as Chair of the Coordination Committee of the Environment for Development Initiative from 2014 to 2019, which has a research center in Addis Ababa that works on water resource management issues in Ethiopia. From 2018 to 2022, he participated in the University of Manchester's FutureDams project, which has an active research collaboration with the Eastern Nile Technical Regional Office. In 2021–2023, he has advised the United Arab Emirates on Nile water management issues.

Jim Hall is a co-investigator of the Foreign, Commonwealth and Development Office (FCDO)-funded REACH (Improving water security for the poor) programme and the GCRF Water Security and Sustainable Development Hub. Both of these programmes conduct research in Ethiopia in partnership with Ethiopian institutions.

Anna Murgatroyd has conducted research for the GCRF Water Security and Sustainable Development Hub, which is in partnership with Ethiopian institutions.

Kevin Wheeler has provided consulting services related to model development, Nile water management, and understanding the implications of the Grand Ethiopian Renaissance Dam since 2012 for the Nile Basin Initiative, the World Bank, Stockholm International Water Institute, the Water Resources Research Institute, Egypt, and the United Nations Environment Programme. Since 2012 he has had ongoing academic collaborations with University of Khartoum, University of Addis Ababa, Cairo University, and Ain Shams University. In 2014–2015 he participated as a member of the International, Non-partisan Eastern Nile Working Group, convened by the Massachusetts Institute of Technology. Contributions to research include GIZ on behalf of the German Federal Foreign Office and Oxford Martin School Programme on Transboundary Resource Management.

Data cannot be made publicly available; readers should contact the corresponding author for details.