The Navigable Frontier Canal of the African Rift

Abstract

This Multipurpose Canal is one of the biggest infrastructure projects for Africa in the 21st Century. It aims at linking 3 river basins, Nile, Congo, and Zambezi. It will eventually connect all the Great Lakes of the African Rift: Albert, Edward, Kivu, Tanganyika and Malawi, as well as Victoria.

It is Multipurpose, as it will involve Hydropower, Navigation, as well as Security and Peace aspects.

Here, are described only Navigation aspects.

This project presents many challenges to the designer, mostly because of the huge difference of elevation between the lakes, and of the divides to climb, reaching up to 2 000 m between the Nile and Congo basins.

The Report presents the design criteria, as well as tentative alignments, possible canalisation steps and techniques to be used over this route, in particular Waterslope, Blue Wave and Pulsating canalisation.

This project is promoted by CSP-REGLA, a Congolese NGO. The first step, Integrated Development of Semliki River, is being studied by the Ministry of Regional Integration of DRC. More details are given about this particular project.

1 Introduction

This Multipurpose Canal is one of the biggest infrastructure projects for Africa in the 21st Century. It aims at linking 3 river basins, Nile, Congo, and Zambezi. It will eventually connect all the Great Lakes of the African Rift: Albert, Edward, Kivu, Tanganyika and Malawi, as well as Victoria.

It is Multipurpose, as it will involve Hydropower, Navigation, as well as Security and Peace aspects.

Here, only Navigation aspects are discussed.

This project presents many challenges to the designer, mostly because of the huge difference of elevation between the lakes, and of the divides to climb, reaching up to 2 000-m above sea level (masl) between the Nile and Congo basins.

Here are presented tentative alignments, as well as the design criteria, possible canalisation steps and techniques to be used over this route, in particular Waterslope, Blue Wave and Pulsating canalisation. These 3 techniques have been invented by members of PIANC and presented in the Bulletin of the organisation of their time.

This project is promoted by CSP-REGLA, a Congolese NGO, and is known under the local name “Sula ya Amani” (the Face of Peace). The first step, Integrated Development of Semliki River, is being studied by the Ministry of Regional Integration of DRC. More details are given about this particular project.

2 The Proposed Route of the Project

2.1 A 3 000 km Link Towards the Ocean

The route shall start in the Nile Basin, on Lake Albert, elevation 616 (masl), reach Lake Edward at elevation 913 through the Semliki River, and join with Lake Victoria, elevation 1 137, through Lake George, elevation 915, and a divide canal at elevation 1 197.

To cross into the Congo Basin, the route will leave Lake Edward and pass along 2 active volcanoes, with a divide reach around 2 068 m, then descend onto Lake Kivu, elevation 1 463. From there, it will reach Lake Tanganyika, elevation 768, through Ruzizi River and connect with the Congo-Lualaba, elevation 548, through Lukuga River.

Crossing into Zambezi Basin can be through Tanzania (divide at 1 430 m) or Zambia (divide at 1 630 m), to reach Lake Malawi, elevation 477. Then, through Lake Malombe and Shire River, the route will reach Zambezi River, elevation 28, and finally the Indian Ocean.

2.2 Nile Basin

2.2.1 Lake Albert-Lake Edward Link

Navigation on Lake Albert is a classical Lake navigation, with short waves and possible fetch effect when the wind blows in the direction of the main axis of this body of water, measuring some 150 km by 40 km.

The project shall start, from North to South, at the border between RDC and Uganda, in the middle of the lake, at the tri-point between RDC, Northern and Northwestern Regions of Uganda, approximately 2°09′14.96N 31°18′21.97. From there, one branch of Victoria Nile is some 8 km away, in Uganda, the start of the Albert Nile.

Elevation of Lake Albert is given by Google Earth at 616 masl.

Along the direct route, navigation on Lake Albert is 144 km to reach the opposite shore, at the point where it is proposed to locate the outlet of the Navigation Canal of the Semliki.

This Navigation Canal is designed to avoid the numerous and tight bends of the Semliki found in the part which constitutes the border with Uganda. From the Lake, it would go up to opposite Bundibundu, the last RDC village on the right bank of the Semliki, a stretch of some 68 km. At this place, elevation is given around 674 masl, so slope/gradient is practically 1 m/km.

To compensate for this fall of 58 m, a succession of 6 locks could be devised, while it is also feasible to use a Blue Wave system, with one gate every 500 m. Such system would avoid an elevated waterway in the middle of a flat plain, or 10 m deep cuts.

Upstream from the border, the river itself will be developed for navigation, with a succession of dams, sited so as to damage to the least extent the virgin nature of the flood plain, which is included into the Virunga National Park up to Lake Edward, elevation 913 masl.

The proposed solution to overcome the 230m+ difference of elevation between Bundibundu and Lake Edward is studied later in this paper. It has to be noted that up to Lake Edward the route is fully in RDC territory.

2.2.2 Lake Edward-Lake Victoria Link

From Lake Edward, another link could be developed towards Lake Victoria. This could be a great interest for Uganda tourism, as it is fully in Ugandan territory and would pass through an area already developed for tourism, the Kazinga channel, connecting Lake Edward to Lake George.

From Lake George, 915 masl, the route would climb the hill up to 1197 masl, thanks to 2 or 3 Waterslopes.

From this divide point, the descent towards Lake Victoria will be easy, only 62 m fall in 175 km, possibly using Blue Waves, 3 gates every 3 000 m.

The working of Waterslopes and Blue waves is explained later in this paper.

2.3 Congo Basin

From Lake Edward to Lake Kivu, and hence to Lake Tanganyika, the proposed route will fall into the Congo Basin. It is the part which poses the highest challenge to the designer, with a divide around 2068 masl, a 1 150 m difference to climb!

2.3.1 Lake Edward-Lake Kivu Link

This link can be designed either on the East side of the volcanoes, or on the West side.

Here is one possible alignment of the East version, North being to the right of this picture:

This route would be exposed to lava from Nyiragongo, but at the same time it may help to divert the flow of lava from inhabited areas.

The West version can be better described in two parts:

The rising part from Lake Edward till the divide, and the divide reach plus the descending part towards Lake Kivu.

Rising from 912 masl on Lake Edward, it takes 50 km to reach elevation 1050 m. From there, it takes 33 km to reach the divide at 2068 masl, with an average slope less than 4%, which leaves the possibility to site there a succession of 13 Waterslopes, 2 km long each, climbing 80 m each. The last 160 m will be tricky to arrange, because the slope of the natural terrain there is more like a cliff, and the structure will have to be sited askew.

On the first 50 km it is possible to use either 8 locks, or a long Blue Wave with some 140 gates, or a combination of both.

From the divide, there will be a 11 km long divide reach at elevation 2068 masl, then a more or less regular descent towards Lake Kivu, down to elevation 1461 masl, according to Google Earth measurements.

This 600 m fall will be compensated by Waterslopes, 2 with a 2.3% slope, between 2068 and 1908 masl over 7 km, and 7 with a 4% slope over 14 km.

This route would be exposed to lava from the Nyamuragira. The Waterslopes would be sited out of the historical domain of this very fluid lava, but the divide reach is likely to be destroyed now and then. Since it is only a canal, it will be easy to reconstruct, on average something like every 15 years.

Another point of interest is that this route follows in many places the limits of the Virunga National Park, except for the divide reach. This would facilitate the protection from encroachment and bandits.

This combined route is 116 km long, 18 km shorter than the East version.

Thanks to the use of Waterslopes, the use of water will be minimal, and transfers from one basin to the other are likely to be negligible.

Both versions fall entirely in RDC territory.

2.3.2 Lake Kivu-Lake Tanganyika Link

To Reach Lake Tanganyika, waters of Lake Kivu follow a river, the Ruzizi, which marks the border between RDC, Rwanda and Burundi. This Ruzizi River is 159 km long and has a head of 645 m.

This river is already being developed for hydropower, thanks to treaties, and navigation could use this diplomatic way in order to be established. Thus we shall not study this link in detail, since it will be a matter of diplomacy rather than technique.

Nevertheless, a number of short-circuit can be devised, the river being replaced by a lateral canal in RDC territory. In particular, a large part of the 110 km where the river makes the border with Burundi is easy to canalise, the terrain being rather flat. A good location for a 28 m dam has been found at 83.7 km, and could provide a 17 km navigable reach. Further, the final 10 km could be short-circuited, since it falls entirely in Burundi.

As far as the upper part is concerned, it could be developed with a combination of locks and Waterslopes, the final design being subject to the negotiations between RDC and Rwanda.

2.4 Zambezi Basin

Passing from Congo basin to Zambezi basin will also be challenging, with again a 1 150 m difference to overcome. Two routes can be envisioned, one through Zambian territory, the other through Tanzania.

Through the Zambian territory, there will be a first 900 m to climb up to the divide, then a slow descent down to the Malawi border, and a 1 150 m descent towards Lake Malawi, as can be seen in the profile.

This route would have the advantage of not impacting the virgin character of Lake Rukwa, which is recommended to be protected from invasion of exogenous species.

The other route would be through Tanzanian territory, with 2 mountain ranges to cross and using Lake Rukwa over nearly 150 km. Although it would seem more obvious, ecological considerations and the 4 slopes to negotiate, rather than 2 in the other option, put this route as a worse solution.

Lake Malawi, 465 masl, is fully in the Zambezi basin, yet reaching the Zambezi River proper is not an easy task, because the river leading to it does not always flow year-round, and is so rocky and narrow that finding a navigable alignment will be difficult. Further, Malawi government has deferred its plans to develop an inland port on its territory, on the banks of the Shire River, some 100 km from the Zambezi River.

Similarly, the Mozambique government does not have plans to create a major harbour at the mouth of the Zambezi. Thus, inland vessels would be unable to unload their cargo in sea ships there. However, to navigate these big lakes, the vessels must be of a stronger scantling than a pure inland ship, and most of the vessels reaching the mouth of the Zambezi would also be able to navigate along the coast up to the existing ports of Beira and Quelimane, or even up to Maputo and Nacala, and the existing small port of Chinde would cater for the other, smaller boats. However, such big developments shall be done with the full support of the governments, and that will be obtained only when a convincing development of traffic is seen over the stretch Lake Albert-Lake Tanganyika.

2.5 Lukuga/Lualaba Development

Another route was studied, the link between Lake Tanganyika and the Congo River. The Lualaba is the name given to that part of the Congo River upstream Kisangani, coming from Katanga. Taking into consideration the mineral riches of Katanga, it would seem sensible to develop a link between Lake Tanganyika and Katanga, to complement the rail link towards the Indian Ocean. At the same time, it would enable to draw a navigable connection with the Congo River proper across the various rapids, and put Kinshasa in contact with the Navigable Frontier Canal of the African Rift. At a later stage, if navigation through the Inga cataracts is made possible, it would be the seed of a cross-continent navigable connection from the Atlantic Ocean to the Indian Ocean.

As a first step, it means to render the Lukuga River navigable. This river is the outlet of Lake Tanganyika, and its flow is very variable, depending on the rains over the Lake catchment. A variation of 11.6 m of the lake level is on record, since the times of Stanley and Livingstone, when the Lake was endorheic. Thus, the first reach of this project will be a “divide reach” starting in the Lake, with a way to overcome a huge variation.

Due to the topography and the position of human settlements, this divide reach may be broken in 2 parts. The higher may be used only when Lake levels are high, to protect human settlements and the existing railway line.

Assuming existing Lake level of 768 masl as on Google Earth¹ and the entrance of Lukuga at 333 km from the Lualaba, a lock & dam would be sited at 303 km and be level with the lowest elevation of the Lake, assumed at 766 masl. The second lock & dam would be sited at 323 km, and compensate for every variation, between 766 and 777 masl. The bottom of both parts would be sited at 762 masl, so as to provide 4 m depth even at low water.

The beam of the largest vessel on the Lake being 17 m, locks would be 18 m wide, and length could be 100 m.

Below this divide, the river can be divided in 3 parts, as far as slope is concerned:

• From 230 km to 303 km, a little less than 1 m per kilometer (67 m in 73 km)

• From 165 km to 230 km, about 1.6 m per kilometer (101 m in 65.5 km)

• From the Origin in the Lualaba till 165, a very slight slope, about 0.33 m per kilometer (51 m in 163.5 km). The last 86 km are even flatter, since the Lukuga only descends 19 m in 80 km (0.24 m/km).

The discharge is extremely variable, from 23 to 2 300 m³/s, and a long-term average has not been computed. It is assumed to be in the range of 300 m³/s below Niemba, where a large tributary meets the Lukuga, and 200 m³/s out of the Lake.

The first part is strewn with rapids, and the proximity of the railway line precludes the use of long reservoirs. We find the same problem as on the divide reach. For all tentative dams up to the divide reach, the great variability of the flow from year to year makes it difficult to make profitable a hydroelectric development upstream of Niemba, with floods one year, minute flow the next, while falls can be only no more than 10 or 13 m because of the railway. To pass the rapids, Blue Waves may be used. An example is shown later in this paper. Small lock & dams may also be used. Insufficient knowledge of the terrain forbids to be more precise.

The second part would be a very good case for hydroelectric development, with 116 m fall in 64 km. 6 locations have been identified, with heads varying between 10 and 25 m. These rather low heads have been selected so that dams be not too wide, inundated area be restricted and that they can be passed by single locks.

As for the dam at 716 masl (width = 815 m), it is set at the highest point so as not to flood the town of Niemba. Its reservoir extends into the first part, and it will even be necessary to plan for a Blue Wave on the last kilometers upstream, or else a deep rock cut so that the boats can reach a possible next dam. The presence of the railway line would make it necessary to lower the level of this next dam, or to rebuild the line over several tens of kilometres.

Below 170 km, the plain is too flat to enable siting large dams. It will be studied at a later stage whether a free-flowing navigation would be possible, or whether some navigation barrages are necessary.

3 Proposed Techniques to Be Used in This Project

3.1 Design of Low-Head Developments

To render navigable an area of rapids, 3 solutions can be proposed depending on the height of the fall of the rapids:

• For a fall of approximately 1 m, a secure pass solution, with a length of less than 500 m.

• For a fall of 2 to 3 m, a Blue Wave, with a length that can exceed 1 km.

• For a higher fall, 5 m and above, a lock.

3.1.1 Secure Pass

The secure pass is well known in Europe as the “Passe Marinière” and was used to cross water mills on rivers. It then required a means of level control, to guarantee the fall used by the mills. It has disappeared, replaced little by little by locks. Two remain in Prague. It has been applied in particular to the Zinga cataract, on the Ubangui, at the border between the CAR and the DRC, and allows the continuity of navigation between Kinshasa or Brazzaville, and Bangui.

A possible application would be at 290 km of Lukuga: a safe route may be devised, where the weak current flows without preventing the ascent. Most of the current follows the cataract, sucked up by the steepest fall.

3.1.2 Blue Wave

The Blue Wave is a patent revealed to the general public by an article in the PIANC Bulletin in 2001. It was there that it was found by the team of Sula Ya Amani. Depending on the project boat and the project speed, the length of each intermediate reach will be calculated for each location, here probably around 200 m. A notable advantage of the Blue Wave, its walls can even be curved and do not have to be straight, when it reuses an old channel of the rapids for example. A possible application would be right at the end of the dividing reach, but the precision of Google Earth’s Digital Terrain Model does not allow exact values to be given: the water at the top of certain rapids is lower than downstream. These will only be simple examples.

At the downstream end of the divide reach, a lock is necessary, because the fall can exceed 6 m. 766 masl is the lowest known altitude of Lake Tanganyika, while the other reach of the “divide reach” serves to compensate for the fall during high waters, up to 777 masl. If we take the “Amani”, the largest boat sailing on Tanganyika (90 × 17 m), as a reference boat, locks should be 18 m wide and 100 m long.

Downstream of the divide reach, falls are much lower, and we enter the Blue Wave area of expertise. The Blue Wave itself is made up of a number of cells of about 200/500 m each. As a lot of water is available, we can use the upstream and downstream cells as buffers, and save a cell, both downhill and uphill. Here, to cross the first rapids, the canal is 800 m, we have therefore only drawn 4 cells, each 200 m long, whereas in a canal we would need 5, or even 6. Another paper in this Congress gives more details on Blue Wave, with an example.

Below are the successive levels that a boat would encounter when descending this Blue Wave. In blue (figures and lines), these are the starting levels, the 2 m drop being divided into 3 stages of 66 cm. In green, we find the figures and levels of equilibrium, when we open the valve between 2 cells. It is 33 cm lower than the upper reach, and 33 cm higher than the next cell. Here, 765.00 masl for the first cell.

The boat having penetrated from one cell to the other, the intermediate door is closed, and the valve is opened towards the 3rd cell. There is a 1 m drop here, with equilibrium at 764.50. The boat enters the 3rd cell. The u/s door of the 3rd cell closes, the valves feeding the 4th cell, 1.17 m lower, open and the levels equalize at 763.915; the d/s door open and the boat moves forward in the 4th cell. The door closes behind her, and a bottom valve open to lower the level of this cell to the downstream level, that is to say 763.33. The downstream door then opens, and the boat can exit the Blue Wave.

On the ascent, we start from altitude 763.33, in yellow this time. By equalizing with the level of the 3rd cell (green), we reach 763.6225 (yellow), and the boat can pass in the 3rd cell. By equalizing with the level of the 2nd cell (green), we reach 764.0613 (yellow), and the boat can pass in the 2nd cell. By equalizing with the level of the 1st cell (green), we reach 764.53 (yellow), and the boat can pass in the 1st cell. Once the door is closed behind her, an upstream valve is opened to bring the level of the cell back to that of the u/s forebay, in order to be able to open the door and allow the boat to enter this u/s forebay.

The bottom of each cell should be 4 m below the lowest level found in the cell, either blue figure or yellow figure. Each cell is as wide as a normal canal, thus there is the same amount of earthworks² than a bypass canal with a lock.

By integrating the average of a possible wait, the passage with a lock would last 2 h, and only 1h32 with the Blue Wave.

3.1.3 Lock

This technique is well proven and well known. It has the advantage of experience and reliability. Thus, we shall not give details about it.

Its domain ranges between 5 m and 30 m of head and is best applicable between 10 m to 25 m. Lower falls are costly for a difference of height manageable by other techniques, and higher falls are in the domain of high-head devices.

3.2 Design of High-Head Developments

For high-head developments, there are 4 techniques available, which will be studied here, keeping in view the extreme difference of levels to be covered, up to 1 150 m:

• Locks, single or in succession

• Inclined planes

• Elevators

• Waterslopes

3.2.1 Lock

The highest lock is located in the former USSR. Today, it is in Kazakhstan, the lock of Ust-Kamenogorsk, renamed Oskemen, on the Irtych. It has a drop of about 40 m. The 3-Gorges dam locking system in China is higher, 118 m, but it is made of 2 lines of 5 cascading locks. Itaipu, in South America, has the same drop, but there is no lock yet. In Brazil, the Tucurui dam (73 m drop) is crossed by 2 locks of 35 m drop, at each end of an intermediate reach.

It is usually considered that the cheapest lock per metre of head is that of 25 m head. Combination of locks enable to cover any difference of height, but strong earthworks are necessary when the terrain is gently sloping.

3.2.2 Incline Planes

The domain of Incline Planes ranges from 35 to 150 m+. Examples are found starting from 13 m, like in Elblag (Poland, 5 units totaling 99 m), with the highest being 110 m at Krasnoiarsk (Russia), which has 2 railway tracks of opposite slopes connected by a huge turntable. The caisson is a self-propelled system.

There are 2 types, longitudinal (Ronquières, 68 m, and Krasnoiarsk) or transversal (St Louis-Arzviller, 44.5 m). Some longitudinal carry the boats in the dry (Biwako, Japan, and Elblag), which is difficult for boats carrying goods. Most have counterweights, but Krasnoiarsk is self-propelled, and could have any length. All examples are in straight line, which makes it at times difficult to place in the existing terrain.

Their disadvantage is the weight of the steel caisson containing the water on which the boat floats, and the difficult foundations of the steel tracks, which carry the water, the caisson and the possible counterweight. The pressure on the underlying ground is extremely high and concentrated. It is usually a very costly proposition.

3.2.3 Elevators

The same problem, cost, is found with elevators, because here the weight is even more concentrated, due to the weight of the structure which guides the caisson full of water. It took 20 years to finish Strepy-Thieu, due to problems of cost and probably of foundations.

The highest on record is at the 3 Gorges dam, 118 m. The difficulty there was to cope with the variation of the upstream reservoir. A sensible solution has been found.

Another difficulty is to find a location where the head is sufficiently concentrated, and this usually obliges to make a deep cut downstream and provide a canal-bridge upstream, both adding to the cost.

3.2.4 Waterslopes

The problem of difficult foundations does not exist with Waterslopes³.

Their principle is to push a wedge of water, on which a boat can float, using a Shield, sort of movable dam attached to propelled wheels, powered by electric motors. This is called Shield (“Bouclier”) or Pusher.

As can be seen on the figure, the weight of the wedge of water is spread on a long surface, and the most weight is found under the shield and its pusher, which must be heavy for adherence (grip) purposes.

As with self-propelled inclined plane, the maximum elevation to cover is practically infinite, 100 m, 500 m, etc. For traffic considerations, a head of 1 150 m will be compensated by 14 Waterslopes, some 80 m each, rather than by a single one of 1 150 m of elevation. The big one would be totally feasible, and much cheaper, but would take some 8 h to climb, so only 3 boats per day could pass, rather than 36 with 14 Waterslopes. Besides, it is unlikely to find a 28 km slope regular and straight. 14 Waterslopes would better adjust to the terrain.

The necessary power is not linked to the difference of elevation, only to the mass of water to move upwards, which itself is linked to the width, the length and draught of the boat, and the slope.

It would seem that, with a slope of 4% and a speed of 1 m/s, Waterslopes should better be 2 km long, with a head of 80 m each. For a width of 18 m and a draught of 2.8 m, the necessary power for the pusher is below 5 000 kW, less than a big electric locomotive. This was found feasible on most of the routes analyzed. At times, a slope of 5% is necessary, at times only 2.3%, as in the descent towards Lake Kivu, and the length may be tailored to what the terrain requires.

Although it is best with a straight alignment, a Waterslope has some possibility to provide curves and counter-curves down to a radius of 1 300 m approximately, for 95 m long boats, when, as proposed, boats have a beam of 17 m and the Waterslope is 18 m wide. This flexibility may be very useful to find alignments with the least excavations. Thus, locating a set of Waterslopes will be rather easy on most of the routes experienced by the Frontier Canal.

Here are pictures of Montech Waterslope, which worked for 35 years until traffic became too small to run it. This 6 m wide structure used 2 locomotives on tyres, with a 1 500 kW diesel-electric propulsion for 3% slope.

4 Application to Semliki River Development

4.1 General Concept

The main issue is the steep fall of about 160 m, concentrated over a few kilometres. A preliminary design for a hydroelectric dam has been found at 105 km, with a reservoir extending over 34 km. But how can navigation pass it? The second issue is that this part of the river falls entirely within the limits of the Virunga National Park, thus it is necessary to tamper to the least extent with the virgin forest. This led to multiply dams, to less inundate the Park. The third issue was the large variation of Lake Edward over the ages, with an extreme of 15 m, which may oblige to deepen the upper segment of the river, in order navigation be possible even at low stages.

4.2 Navigating a 162 m Fall

The 162 m drop on the Semliki between the headrace and the downstream compensation reservoir is currently the highest navigable height difference studied in the world. Overcoming it is possible in a number of ways, with 7 or 8 high head locks for instance. Here is only shown the solution using 3 Waterslopes.

The trick is to use a contour canal, following elevation 899 masl, branching out from the reservoir at the 105 km dam. This 324 m² canal, 36 m wide, 9 m deep, will be navigable for about 7 km, and this headrace doubles as a feeder canal to bring the flow of the river to the 3.3 km underground penstocks feeding the turbines.

At the end of this contour canal, there would be a first Waterslope, from 899 to 850 masl, then a 1.4 km embankment at 850 masl to cross one tributary of the Semliki, the Butawu, then a second, bended Waterslope from 850 to 794 masl, which follows more or less the border of the Virunga National Park, then 2 options, depending on whether the Park insists on protecting the Luusilubi River, in which case the third Waterslope will be again bended, or this river can be developed for navigation, with a straight Waterslope leading to it.

By the way, this Luusilubi is the highest source of River Nile, vying with the Butawu for this title, these brooks starting above elevation 4 650 m.

4.3 Developing the River Below the Dam

Attempts were made to design barrages between elevation 737 and the border. It seems that eight are needed, avoiding a ninth, too close to the border and which would risk, by reducing the solid flow, modifying the course of the border river, which could be the source of many territorial conflicts.

The highest barrage in this provisional design is only 12 m high, most are 7 to 10 m high, and 2 are only 4 m high. This was necessary, because the land is very flat, and 1 m more of head increases the flooded area by several tens of hectares. Despite the low drop, it is difficult to imagine the use of Blue Waves; the river is, in fact, very wide and there are no rapids on which a Blue Wave could be grafted. Thus, dams would be necessary in any case.

Due to an imprecise knowledge of the topography and the small discharge of the Semliki (200 m³/s), the possibility of a Pulsating Canalisation has not been thought possible either.

For navigation, there are a large number of steep curves, radius below 200 m, and if not covered by reservoirs, they will have to be short-circuited.

4.4 Developing the River Above the Dam

Between Lake Edward and the proposed dam, only 2 barrages are necessary: one to cope with Lake Edward variations, the other to serve as the end of the “divide reach” when Lake Edward is extremely low. There is evidence of historical variations of Lake Edward down to 15 m lower than present. However, it was most probably endorheic in those dry years, so the lowest navigable level will be said to be 908 masl only. Presently, Google Earth gives an elevation of 913 for Lake Edward, and the highest high level will be said to be 917 masl. Gauges around the lake give figures some 6 m higher, but since our sole reference is Google Earth, we shall use its values.

Both parts of the divide reach may only flood land likely to be covered by an exceptional flood. An interesting location is proposed for the first barrage, with a controlling valve culminating at 917 masl, in case Lake Edward rises to the extreme and floods all the reservoir, but used mostly around 913 masl. A lock would compensate for these variations and lower vessels at elevation 908 masl. This reach, as well as the second, would be dug at 904 masl so that navigation is possible even in the driest year. In such case, the lock would not be used.

The second barrage would have a head of only 9 m and be dredged at 904 masl, but to avoid inundating one extra square kilometre the 908 masl reach will have to be shortened by 4 km and the 899 masl reservoir extended accordingly by dredging over 2 kms.

In the reservoir, a 4 m channel will have to be provided, marked and maintained.

5 Conclusion

This paper provides the proof of a technical feasibility for the Navigable Frontier Canal of the African Rift, as well as its junction both with Lake Victoria and with the Congo (Lualaba) River.

Its hydroelectrical potential may be sufficient to fund the cost of the first element of this Canal, the Semliki River Development, an International waterway entirely in RDC territory.

It may also be the occasion to test in full scale the Waterslope and Blue Wave techniques, which offer an interesting potential and may render workable new navigations in mountainous terrain. They are further very well adapted to autonomous navigation, prepared for the future.