Water Level Monitoring of Grand Ethiopian Renaissance Dam using SAR data

7 min readJul 23, 2020

The Grand Ethiopian Renaissance Dam (GERD) is a dam on the Blue Nile River in Ethiopia that has been under construction since 2011. It is in the Benishangul-Gumuz Region of Ethiopia, about 15 km east of the border with Sudan. At 6.45 gigawatts, the dam will be the largest hydroelectric power plant in Africa when completed, as well as the seventh largest in the world.

Grand Ethiopian Renaissance Dam Localization.

The GERD project is a major tributary of the world’s longest river the Nile River contributing up to 80% of its water during the rainy season.

The Blue Nile originates at Lake Tana in north-western Ethiopian Highlands and runs for approximately 1,450 km to meet another major tributary the White Nile in Khartoum, Sudan.

Dam characteristics

The zero level of the main dam, the ground level, will be at a height of almost exactly 500m above sea level, corresponding roughly to the level of the river bed of the Blue Nile. Counting from the ground level, the main gravity dam will be 155m tall, 1,780m long and composed of roller-compacted concrete. The crest of the dam will be at a height of 655m above sea level. The outlets of the two powerhouses are below the ground level, the total height of the dam will, therefore, be slightly higher than that of the given height of the dam. In some publications, the main contractor constructing the dam puts forward a number of 170m for the dam height, which might account for the additional depth of the dam below ground level, which would mean 15m of excavations from the basement before filling the dam. The structural volume of the dam will be 10,200,000m³ .

Sentinel 2 RGB Image from May of 2020. (Google Earth Engine)

Supporting the main dam and reservoir will be a curved and 5.2km long and 50m high rock-fill saddle dam. The ground level of the saddle dam is at an elevation of about 600m above sea level. The surface of the saddle dam has a bituminous finish, to keep the interior of the dam dry. The saddle dam will be just 3.3–3.5km away from the border with Sudan, it is much closer to the border than the main dam.

The reservoir behind both dams will have a storage capacity of 74km³ and a surface area of 1,874km² when at full supply level of 640m above sea level. The full supply level is therefore 140m above the ground level of the main dam. Hydropower generation can happen between reservoir levels of 590m , the so-called minimum operating level, and 640m , the full supply level. The live storage volume, usable for power generation between both levels is then 59.2km³ . The first 90m of the height of the dam will be a dead height for the reservoir, leading to a dead storage volume of the reservoir of 14.8km³ .

filling the reservoir

At the beginning of July, the Government of Ethiopia began the process of filling the reservoir with the help of rain water from the region’s wet season.

In the first year, the Gerd will retain 4.9 billion cubic meters (bcm) of water, taking it up to the height of the lowest point on the dam wall, allowing Ethiopia to test the first set of turbines. On average, the total annual flow of the Blue Nile is 49bcm.

In the dry season the lake will recede a bit, allowing for the dam wall to be built up and in the second year a further 13.5bcm will be retained.

By that time, the water level should have reached the second set of turbines, meaning that the flow of water can be managed more deliberately.

Impact on Egypt and Sudan

The precise impact of the dam on the downstream countries is not known. Egypt fears a temporary reduction of water availability due to the filling of the dam and a permanent reduction because of evaporation from the reservoir. Studies indicate that the primary factors which will govern the impacts during the reservoir filling phase include the initial reservoir elevation of the Aswan High Dam, the rainfall that occurs during the filling period and the negotiated arrangement between the three countries. These studies also show that only through close and continuous coordination, the risks of negative impacts can be minimized or eliminated.

The Grand Ethiopian Renaissance Dam could also lead to a permanent lowering of the water level in Lake Nasser if floods are stored instead in Ethiopia. This would reduce the current evaporation of more than 10 billion cubic meters per year, but it would also reduce the ability of the Aswan High Dam to produce hydropower to the tune of a 100 MW loss of generating capacity for a 3 m reduction of the water level. However, the increased storage in Ethiopia can provide a greater buffer to shortages in Sudan and Egypt during years of future drought, if the countries can reach a compromise.

For More:

Could mega-dams kill the mighty River Nile? (an interactive report)

For the 280 million people from 11 countries who live along the banks of the Nile, it symbolises life. For Ethiopia, a…

interactive.aljazeera.com

SAR for Water Level Monitoring

In this case study, we will use SAR images from Sentienel 1 together with a Digital Surface Model obtained by the ALOS PALSAR 2 satellite to monitor the water level at GERD.

The use of SAR data for mapping water bodies has been widely used in several studies. This is due to the way the water interacts with the RADAR waves. Water surfaces are generally smooth and are characterized by very low radar backscatter, but this is affected by wind roughening effects that result in the “erasing” of some water features on radar imagery. The detection of surface waters is based on the difference in roughness between water-covered areas and other types of land cover. The higher the backscatter, the brighter the feature looks on the image and vice versa. As noted above, water features have very low roughness, and thus very low backscatter, meaning they appear as very dark features that contrast strongly with their surroundings.

Sentinel 1 data on Google Earth Engine

The Google Earth Engine plataform facilitates data access and analysis. Let’s compare the water level in the month of June (before the start of the damming of the water) with that of the month of July (after the start). First we get the sentinel 1 images for the defined dates:

Sentinel 1 - VV band Comparation: June 2020 / July 2020.

Alos Palsar DSM

To calculate the elevation and mapping of the water, we will use the digital surface model of the ALOS satellite, with a horizontal resolution of approximately 30 meters. The maximum elevation estimated in the region of analysis was 1863 meters, while the minimum elevation is 488 meters. To facilitate visualization, we create contour lines every 50 meters.

With the data in hand, we will use the supervised classification to separate water from non-water. As features we will use the VV and VH bands of the SAR image together with the digital surface model. We collected samples from both classes and performed the training of the random Forest algorithm. Thus we obtained the mapping of the areas with water in the June and July images.

After obtaining the masks of the bodies of water, it is possible to calculate the size of the flooded area in June and July:

Finally, we select an area close to the two dams to check up to what elevation the water has reached:

Conclusion

The Grand Ethiopian Renaissance Dam is a huge project on the Blue Nile that has its benefits and problems. Through images from the Sentinel 1 Satellite, we can see that filling of its reservoir has already started. By July 10, an area of approximately 30 km² had been flooded, raising the water level by 27 meters. The reservoir will continue filling until it reaches its desired level. The Google Earth Engine can be a good tool for future analyzes of the impact of this dam in the region of its construction and also downstream, where there is a huge Sudanese and Egyptian population.

Thanks!