The downstream impact of the first and second filling of the Grand Ethiopian Renaissance Dam

Water conflicts arise from geostrategic factors that hide behind visible ones such as the construction of dams. In the case of the Grand Ethiopian Renaissance Dam (GERD), it will have a significant impact on development in Ethiopia, but its filling has worried downstream countries about altering the flow of the Nile and associated ecosystems. In this study, several high spatiotemporal resolution remote sensing products were used on the basis of artificial intelligence in Google Earth Engine. The results show that the two first filling phases had no effect on the reservoirs of the dams in Sudan and Egypt, nor on the vegetation cover. On the contrary, significant reserves of water have been stored in response to unusual floods in the White Nile, and the recent trend of increasing vegetation cover has not been affected likely due to groundwater contributions and judicious anticipation, and the provision of resources for food security. In general, the parties must take long-term collaborative measures to ensure effective management and reduce waste, especially in the upcoming filling.


Introduction
Egypt has long dominated the resources of the Nile Basin under a strategic umbrella that hampers any competition for water supply. However, Ethiopia has recently challenged the status quo through several measures, agreements and initiatives, the most recent of which was the announcement in 2011 of its intention to build the Grand Ethiopian Renaissance Dam (GERD) on the Blue Nile about 40 kilometres East of Sudan (Tekuya 2020). This has led to renewed geopolitical tensions and water-related conflicts that have raged for three decades, despite appeasement efforts under international auspices that seek to guarantee each country's share and protect its water security (Nasr and Neef 2016;Turhan 2021). The announcement came without prior notice or negotiations, hence the subsequent aggressive verbal response from Egypt, which expressed serious concerns about reducing its water share given the potential control capabilities of the GERD, considered by Egypt as a serious threat to its water, power, and agricultural supply (Ashour et al. 2019;Makonye 2021).
In geopolitics, water conflicts stem from indirect factors such as unequal power relations and water ambitions, and their spark is ignited by direct factors such as building dams (Zeitoun and Warner 2006). The close connection between water and geopolitics arises throughout history when water scarcity becomes a security issue and a matter of existence (Salhi et al. 2022b). Each state reveals and argues its point of view and defends its interests to the end. Ethiopia argues it is under severe regional agricultural stress which, coupled with rapid population growth, will worsen water scarcity in coming decades. Despite projected increases in regional precipitation due to climate change, the frequency of hot and dry extremes is also expected to increase due to warming (Coffel et al. 2019;Tariku and Gan 2018). This leads to a high risk of disruption to agriculture, especially since it is mostly rainfed, making yields vulnerable. Recent droughts have severely reduced crop yields, leading to food insecurity (Coffel et al. 2019). As for Egypt, a North African country that lies in the middle of the Sahara desert in an extremely dry climate, the material and moral link with the Nile is 1 3 14 Page 2 of 10 close like a lifeline even before Herodotus describes Egypt as 'the gift of the Nile' in the 4th century BC (Hassan 2007).
It is certain that the GERD will have a significant economic, social and political impact in Ethiopia, but Egypt fears the consequences of this on changing the flow of the Blue Nile and the ecosystems associated with it, not to mention the apprehension associated with management systems during future droughts (Abtew and Dessu 2019). Yet, water scarcity and declining yields could be overcome through effective cooperation to reduce massive water losses (over 90% of the basin's total inflows), and by drawing benefits from the most successful technical and political experiences implemented in other international river basins (Ashour et al. 2019).
The hydrology of the Nile shows a large inter-annual variability of flows and clear geographical and climatic differences. The annual flow varies between 45.6 billion cubic meters (BCM) and 120 BCM with an annual average of 86.5 BCM. The Blue Nile contributes about 55% of this flow, with the remaining 32% and 13% coming from the White Nile and Atbara, respectively (Wheeler et al. 2020). Egypt depends on it to secure 95% of water for all purposes (El-Nashar and Elyamany 2018). Indeed, under previous agreements, Egypt and Sudan were granted most of the water rights without Ethiopia receiving a specific allocation (Kasimbazi 2010).
According to a recent article, Ethiopia planned a shortterm filling scenario (2020-2022) that could lead to a deficit in Egypt, for instance, equivalent to the losses of the current cultivated area of up to 72% (Heggy et al. 2021). However, this assumption has generated deep scientific debates and controversies (Eladawy et al. 2022;Heggy et al. 2022;Wheeler et al. 2022) that identify a research path to be undertaken.
Furthermore, although fair and judicious use is a fundamental principle in the distribution of water resources, its practical application still presents a great challenge, which leads to a dead end in the allocation of water between riparian countries (Onencan and Van de Walle 2018). This is due to the complexity and lack of trust, factors which do not contribute to highlighting the potential benefits of a regional cooperation approach to addressing the fundamental challenges of water diplomacy, nor to promoting a good understanding of river basins as resources of interdependence and integration at all levels that go beyond abstract physical and temporal scales (Al-Saidi and Hefny 2018). Consequently, there is a need to establish an impartial understanding of the evolution of water amounts in the Blue Nile dams over the past few years and to trace the resulting impact on various environmental aspects to help alleviate the congestion associated with water conflict.
Therefore, this study describes the territorial repercussions of the water conflict related to the impoundment of the GERD between the three major riparian countries of the Nile basin (Egypt, Sudan, and Ethiopia) from a geospatial perspective. Between January 2016 and December 2021, we study by high spatiotemporal resolution imagery the extent of the dam lakes in the three countries, and we analyse over the last twenty years (2002-2021) the main environmental and climatic parameters (average and statistical trend of precipitation, NDVI anomaly, variation in reservoir extent, land use/land cover, SPI, SVI and EVI) to discern their trends and the recent implications of filling the GERD.
Several studies have sought to identify the risks and tradeoffs between different riparian objectives, but our study has the advantage of providing insights into ground facts in crucial geopolitical circumstances. It builds on the benefits of remote sensing for extracting water bodies under different conditions and scenarios. It could serve as a basis for future studies to monitor changes in the waters of the Nile basin, even if it is constrained by the unavailability of altimetric data. Truly, given the limitations of remote sensing, altimetric data may indicate the water level based on storage instead of the approximation taken from the surface area of the lakes.

Material and methods
The analysis was based on remote sensing products from different collections at high spatial and temporal resolution using artificial intelligence on Google Earth Engine (Table 1, Fig. 1). The latter is a powerful and transversal open tool to assess climate change anomalies and their  (Salhi et al. 2022b). The evolution of the reservoirs surface of the three riparian countries' main dams between January 2016 and December 2021 was calculated by filtering the collection of Sentinel-2 images (10 m resolution) according to the date and to a given reservoir. Subsequently, the monthly median was calculated and a function to mask cloud cover (less than 1%) was applied from the built-in quality band information (QA60). Afterwards, the Normalized Difference Water Index (NDWI) was computed for the resulting acquisition based on bands B3 and B8 according to equation 1. The validity and sensitivity of Sentinel-2-based NDWI are widely admitted in several previous articles (Cordeiro et al. 2021;Du et al. 2016). All the pixels with an NDWI value greater than 0.3 were extracted to better isolate surfaces with detectable water surfaces (Gu et al. 2007;McFeeters 2013) and the monthly sum of the hydric area were calculated. Finally, the evolution curve of each reservoir was drawn.
Changing land use patterns is of great importance in environmental studies and critical for land use management decision-making (Lebrini et al. 2021). In our case, it was used a landcover map which is a composite of land use/ land cover (LULC) predictions for 10 classes throughout the year to generate a representative snapshot of 2020. It is derived by Esri (Environmental Systems Research Institute) from ESA Sentinel-2 imagery at 10m resolution (Karra et al. 2021) with a recognised global high accuracy (Ding et al. 2022;Mirmazloumi et al. 2022). The assessment of the pixel-based accumulation of NDVI (Normalized Difference Vegetation Index) anomalies over the last nine years (2013-2021) used the 'MYD13A1.006 Aqua Vegetation Indices 16-Day' dataset according to equation 2 (where NIR and R are Near-infrared and Red images, respectively) (Didan 2015). The reference conditions were defined from the first 10 years of data (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012) then the average at the pixel scale was calculated. Afterwards, the latter was subtracted from each of the images in the data collection (from 2013-01-01 to 2021-12-31) and an iteration was applied over the calculated anomaly images to create a collection of cumulative anomalies over time.
Understanding the trend and variability of precipitation and drought characteristics is of paramount importance for Flowchart of the material and methods used in the study water resources management, agricultural planning, and the study of climate change, as indicated by many previous articles (Mallick et al. 2021;Salhi et al. 2022a;Sharma et al. 2022 These tests quantify the dependency between two variables, where the series are the dependent variable and the time is the independent one (Benabdelouahab et al. 2020). The framework of drought risk analysis provides a unified and coherent approach to solving inference and decisionmaking problems under uncertainty due to climate change (Kim and Jehanzaib 2020). Drought is monitored using specific meteorological and vegetation indices, which measure deviations from normal local conditions based on historical distributions (Dai 2011). In our case, meteorological drought monitoring based on the Standardized Precipitation Index (SPI) was assessed between 1981 and 2021. For the monthly SPI, a list was created with a one-month lag between each list entry, starting with the last image and counting down. Then a simple list was created, and month-end dates were mapped on it, counting down in time. The list was latter sorted by date. Subsequently, the monthly sums of the CHIRPS were calculated according to the same time step. After that, statistical bands (mean and standard deviation) were created to be used for the calculation of the SPI. For the 16-day SPI, a list was created with MOD13Q1.006 collection for each 16-day period. The millisecond format of the list has been changed to a normal date format and a time step has been provided. Next, CHIRPS precipitation data were summed for each 16-day MODIS interval. Afterwards, statistics (mean and standard deviation) were calculated for each image to be used for the SPI calculation. Finally, a chart was plotted to show the evolution of each of the two SPIs.
To monitor vegetation drought, EVI (Enhanced Vegetation Index) and SVI (Standardized Vegetation Index) were chosen to be complementary and accurate in studying vegetation condition trends and monitoring areas affected by drought (Alexandridis et al. 2020;Prananda et al. 2020;Vaiopoulos et al. 2004). Both indices were evaluated for the same period (2000-2021) on the basis of MOD13Q1.006 for each 16-day scale using a process well described in the literature (Mahmud et al. 2021).
The changing population estimate used the 'WorldPop Global Project Population Data' dataset (www. world pop. org) which is very accurate in assessing the total population (Bai et al. 2018). A time band (in milliseconds) was created and a linear regression between 2000 and 2050 was established to estimate the evolution of the population density.

The reservoirs of the dams downstream have preserved or improved their surface
The Blue Nile is regulated by five dams which are Aswan High Dam (Egypt), Merowe, Sennar and Roseires (Sudan), and the GERD (Ethiopia). The waters regulated over the last six years cover an average area of 4,714 km 2 (Lake Nasser in Aswan), 421, 13 and 209 km 2 (Merowe, Sennar and Roseires, respectively), and 36 km 2 (GERD).
The first phase of filling the GERD began in the summer of 2020 when, from August, we observe a gradual rise in the water stored up to cover an area of around 190 km 2 (Fig. 2). Afterwards, we observe a slight decline from May 2021 then a significant rise between September and October of the same year (second phase of filling) and the regulated waters stabilizes at around 370 km 2 .
In Sudan, the Sennar dam lake shows little variation due to its low retention capacity. It is waterless during low water periods, while the area of the lake does not exceed 25 to 30 km 2 during high waters. The lakes of the other two dams (Merowe and Roseires) show a similar sequential evolution curve. A period of very low water is observed for the two lakes between June and August, then a significant rise in September-October and a stabilization of maximum water (over 600 and over 400 km 2 for Merowe and Roseires, respectively) between November and March approximately.
In Egypt, the shape of the surface of Lake Nasser is roughly like that of the Sudanian lakes except that the magnitude is much greater. Before the beginning of the filling of the GERD, the maximums managed to exceed 5,500 km 2 against minimums which generally varied around 3,000 km 2 . From November 2019, it is observed a replenishment of the endorheic lakes of Toshka from the release of Lake Nasser. Thus, these lakes, which had a surface area that oscillates between 130 and 10 km 2 , have increased to more than 1,300 km 2 in 2020 and 2,115 km 2 in December 2021. From Figure 2, it is clearly noted that Egypt has managed the situation well by anticipating by one year the filling of the GERD by combining the filling of the lakes Nasser (Aswan) and Toshka where the cumulus rises to more than 7,200 km 2 (2020) and to more than 8,200 km 2 (2021) in full phases of filling the GERD. This anticipation prevented the retreat to less than 4,000 km 2 during the low water period of 2021.
In conclusion, it is observed that the two filling phases of GERD did not affect the normal curve of the extent of the lakes downstream. Contrary to speculation, none of the Sudanese lakes receded after the filling of the GERD, and rather Lake Nasser gained widening and diverted large amounts of water into the lakes of Toshka under wise proactive Egyptian management (Fig. 3).

Vegetation cover was not affected
Basin-scale analysis of environmental factors (vegetation cover, NDVI and NDWI) shows the preponderance of vegetation cover upstream and in the extreme downstream. Contrarywise, between Khartoum and the Delta dominate shrubs and bare soil except on the banks of the Nile (Fig. 4a, e). The Normalized Difference Vegetation Index (NDVI) cumulative anomaly during the last 9 years (2013-2021) compared to the reference period (2002-2012) shows a dominant positive anomaly upstream of the Blue Nile and a dominant negative anomaly at the Delta ( Fig. 4b and 4f). In Ethiopia, however, many areas in the north and centre are devoid or have little vegetation and have shown a negative anomaly. In the Delta, in addition to the dominant negative anomaly in the centre, there is an absolute Mann-Kendall negative NDVI trend at the pixel scale between 2002 and 2021 ( Fig. 4f and 4g). This negative statistical trend (Sen's slope < 0 and α < 0.05) prevails on the banks of the Nile between Aswan and the Delta and around the latter, despite the improvement of the vegetation cover reflected by the positive anomaly of the last 9 years (Fig. 4f).
Mann-Kendall statistical trend of precipitation during the last 41 years is significantly positive (Sen's slope > 0 and α < 0.05) in Sudan and upstream of the Blue Nile. In the latter, a not-significant trend (α > 0.05) is also observed (in central and northern Ethiopia) (Fig. 4d). Considering the importance of precipitation in these areas which exceeds 500mm in Sudan and reach 1,700mm annually in Ethiopia (Fig. 4c), the positive trend observed is expected to bring abundant amounts of water, but also justifies the construction of dams to regulate the resource and protect against the growing risks of flooding.
The population density is very high on the banks of the Nile between Aswan and the Mediterranean (over 1,000 inhab./km 2 ) while it drops to less than 250 upstream of the Blue Nile and around Khartoum then to less than 100 inhab./ km 2 elsewhere (Fig. 4h). The expected average population growth in northeast Africa is 3.2% in Ethiopia, 0.8% in Sudan and 3.5% in Egypt (Clement et al. 2021). This means that by 2050 the population upstream will experience significant growth but especially in the Delta where the density will increase to unprecedented levels. The hydro-power demand would therefore be stressful in the Delta to cover exceptional water and food needs.
It is concluded from what has been mentioned that the filling of the GERD at its current pace has not affected the stock of dams in the downstream countries, nor the most important environmental factors (i.e., vegetation and water) in the Nile Delta. Even though, it was sought to analyse the time series of meteorological and vegetation drought indices in the delta to highlight any hidden impacts or anomalies.
To assess meteorological drought, the Standardized Precipitation Index (SPI) was used over two short periods of accumulation (i.e., 1 month and 16 days) to assess the direct effects of soil moisture change and water flow in streams surrounding the Nile and to characterize the temporal pattern of stream flow and reservoir storage in the Delta. SPI-16 (16 days) for the past 22 years  shows frequent interannual oscillations and positive peaks that occur in different seasons. Before 2016, zero values generally predominate the annual curves and are interspersed with maxima that reach On a larger time scale, SPI-1 (1 month) of the last 41 years  shows that the positive sequence which began in 2016 is exceptional although there were two relatively less important positive sequences in 1985-1988 and 1992-1995 (Fig. 5).
For vegetation drought analysis, the Enhanced Vegetation Index (EVI) was used to assess the greenness of the vegetation due to its good sensitivity. It is an optimized vegetation index designed to enhance the vegetation signal with improved sensitivity in high biomass regions and improved vegetation monitoring through a de-coupling of the canopy background signal and a reduction in atmosphere influences . Over the past 22 years , its average values in the Delta show an annual sequence with two positive peaks (maximum at the beginning of the year) followed by two minimum peaks (Fig. 5). The general trend shows three positive sequences: 2001-2007, 2011-2014 and 2018-2021. In addition, it was studied the Standard Vegetation Index (SVI) which characterizes the vegetation conditions. Low values indicate poor vegetation conditions that could be the result of moisture shortages or extreme temperatures, while high values may reflect ideal growing conditions. During the same period, the three positive sequences mentioned by the EVI index have been observed, the last one continuing since 2018 proves that greenness has improved in the Delta despite the measures taken upstream of the Nile. However, the overall SVI trendline is downward due to the first positive sequence being weak and the last sequence being preceded by historical lows that crossed -0.8. This explains the observed negative NDVI anomaly and the negative Mann-Kendall trend. Yet, it is concluded that the first two filling phases of the GERD did not affect either the water reserves or the vegetation of the Delta.

Discussion
The filling of GERD is still minor, far from the short-term scenario initially envisaged by Ethiopia. The GERD reservoir contains only less than 8.5 BCM (compared to the 18.5 BCM predicted by the second phase of filling). This small storage is not comparable with the planned operating storage of the dam, which stands at 74 BCM with the planned third filling and will retain between 20 and 40 BCM of the Blue Nile inflow on an annual basis until the end of filling the tank (Abou Samra and Ali 2021;Heggy et al. 2021;Wheeler et al. 2020).
Given the high level of Lake Nasser in 2020 and 2021 (often above the 178 m level) in response to an unusual flood in the White Nile, the first and second phases of the filling had no effects on the water supply in Egypt. Consequently, the fluctuations of lake areas downstream in Egypt are not mainly correlated to the Blue Nile, where the GERD is being built. The next phase of filling (i.e. 2022), if Ethiopia proceeds with the scheduled plan (>20 BCM), would result in a drastic change in the land cover of Egypt (Abou Samra and Ali 2021).
The Toshka Lakes are a spillway for flood waters to protect the High Aswan Dam body when its lake reaches flood level (178 m). Apparently, the lakes are low topographic areas relatively to the dam and therefore the water that is discharged into the lakes cannot return to the system. Therefore, Toshka Lakes are reserve storage of Nasser Lake even if both cannot be considered as an integrated system. Egypt tried to keep the Lake Nasser levels at a "safe limit" to support electricity generation during the filling phases while the GERD reservoir will have total storage of 74 BCM which is equivalent to nearly 1.2 times the average annual flow of the Blue Nile at the dam site (Wheeler et al. 2020). The short-term filling is likely to bring Lake Nasser to a minimum operational level for four consecutive years, which may interfere with the maximum allowable reduction estimated to be no more than 5-15% of its normal capacity (safe limit) (Abdelhaleem and Helal 2015). In the meanwhile, the annual uncovered deficit can be partially solved by adjusting the operation of the High Aswan Dam, developing groundwater extraction, and adopting new policies for the cultivation of crops (Heggy et al. 2021). Simulations show that the expected amount of saved water using such strategies could exceed the losses caused by GERD by 12.1 BCM (El-Nashar and Elyamany 2018). Thus, it is important to emphasize that agricultural land in Egypt is largely maintained by groundwater resources, even in the vicinity of the Nile Valley. This also explains the upward trend of the vegetation cover.

Conclusions
Despite the speculations and fears (Eladawy et al. 2022;Heggy et al. 2021Heggy et al. , 2022Wheeler et al. 2022), this study shows that the two phases of filling the GERD (2020 and 2021) had no effect on the extent of the dam reservoirs in Sudan and Egypt, nor on the vegetation cover. On the contrary, Aswan Lake expanded, and large quantities of water were transferred to Toshka Lakes to increase the strategic stock in anticipation of a water crisis that did not occur. In the delta, the upward trend of vegetation cover that has continued since 2018, was not affected, thanks to the Egyptian authorities' anticipation of a possible water crisis and the provision of resources to achieve food security. Therefore, the current bet is based on maintaining the existing balance to spare the region any uncalculated escalation that will lead to the fragmentation of the political situation and the emergence of new regional powers that will have different Looking at the issue of GERD from a purely social, economic, and developmental perspective is inaccurate because it ignores the ability of this major water facility to change the geopolitics of the Middle East and North Africa. It is true that population growth, food security, energy needs, economic and technological development, political fragmentation, and international water laws create concerns about water availability in the short and medium term, but concern about changing regional geopolitical balances is the central driver of conflicts over resources (Al-Ansari et al. 2018;Vesco et al. 2020).
The lack of confidence and transparency in managing the issue between Ethiopia and Egypt has fuelled the conflict between them and turned it into a serious threat of using military force without the world powers being able to calm the situation and provide a solution that satisfies both parties. To mitigate this, there is a need for mediation that can bring the parties to the negotiating table by providing specific incentives for cooperation. In addition, all parties must take wise measures and a long-term strategic plan based on comprehensive development to ensure efficient water management and reduce water waste.
Economic prosperity is a function of the balance of geopolitical interests, which are determined not only by the resources of a country but also by the resources of its neighbours (Adhvaryu et al. 2021). Future structural and environmental assessments suggest that the problem of water scarcity will worsen in the lower Nile (Ashour et al. 2019;Coffel et al. 2019), and mutual intransigence will not address this but will lead to decisions that will have intertwined and unpredictable consequences that will further complicate the problem.
Acknowledgements This research received no external funding.

Conflict of interest
The authors are not affiliated with any organization with a direct or indirect financial interest in the subject matter discussed in the manuscript and they have no known competing interests or personal relationships that could have appeared to influence the work reported in this paper.