Version classiqueVersion mobile
OpenEdition Books

Avenir du fleuve Niger

Jérome Marie
Pierre Morand
Hamady N’Djim

Synopsis and Recommendations

The situation and how it is evolving

Texte intégral


How Mali’s society and its requirements are evolving

1This review of Mali’s population today, how it is evolving, and how it will probably evolve through 2025, is based on information supplied in the Mali 2025 National Outlook Study (NOS, 2003).

Mali’s population, how it is evolving, and how it will probably evolve by 2025

2Mali’s 9.8 million inhabitants (by 1998 census figures) are spread unevenly. In 1987, 65% of the population lived on 25% of the land (in the south). This worsened in 1998, when only 30% of the territory was home to 91% of the resident population. Mali’s average population density grew from 6.2 inhabitants per km2 in 1987 to 7.9 inhabitants per km2 in 1998, with sharp contrasts from one region to another. The Timbuktu region, for instance, counts less than 1 inhabitant per km2. Segu, on the other hand, counted 25.9 inhabitants per km2 in 1998. Lastly, the rate of natural increase stood at 3.7% between 1976 and 1987, and remained above the 3% mark from 1987 to 1998. This is high: excluding migration, it means that the population is doubling every 24 years. Extensive emigration, however, relieves some of the pressure and puts Mali’s population increase in a relatively moderate bracket. The resident population only increased 1.8% a year on average from 1976 to 1987, and 2.3% from 1987 to 1998. The average annual increase over this last period was contrasted: the urban population rose 4.5% but the rural population only 1.2%.

3Different scenarios suggest that Mali will count between 16 and 20 million inhabitants in 2025. The rural population should shrink from 73% today to only 51% of the total (i.e. to about 9 million). Bamako should see the sharpest surge (its population should double). Other cities in northern Mali should see sharp population increases as well. However, in absolute terms, population concentration should be the highest in and around Bamako and Segu-Mopti (which will count 2.9 million and 2.4 million inhabitants respectively). This Bamako-Segu-Mopti triangle, in other words, will be home to three-quarters of Mali’s urban population.

How food requirements should evolve through 2025

4Malians consume cereal mainly (204 kg per person and per year), along with limited amounts of meat, fish and vegetables. City dwellers consume a little less cereal (180 kg per person and per year) but more meat, fish and vegetables than the average (the population is still predominantly rural).

5By 2025, Mali will have to produce between four and five million tonnes of cereal a year to keep up with its growing population. With the exception of cereal (“only” 40% of which will be consumed in cities), urban markets will absorb the biggest share of the fish, meat, fruit and vegetable production in 2025 (55 to 75%, compared to less than half in 1995). Consumption of domestically-produced food in cities is expected to rise 3.5 times faster than that in rural environments over the period from 1995 to 2025.

6Cereal, fish, fruit, vegetable and red-meat production, it follows, will have to double – or triple – by 2025. Farming more land will not be enough. Improving yields, it follows, is the key.

Table 2 – Food consumption in Mali (in tonnes)

Table 2 – Food consumption in Mali (in tonnes)

Source: Based on 1988/97 DNSI consumption budget survey

How energy requirements should evolve through 2025

7Traditional sources of energy (wood and charcoal, basically) are still predominant (accounting for about 90% of the total). Vehicle fuel and household gas also account for fairly significant portions of Mali’s energy consumption.

8Electric power – i.e. the source of energy that has a direct relationship with river management – is still relatively marginal, but growing fast. Power production and consumption are evolving pretty much in tandem, making Mali practically self-sufficient in electric power. It neither imported nor exported electricity until 2002. This has since changed: Mali is not exporting power from a balance-of-trade angle, but nonetheless supplying Senegal and Mauritania from Manantali.

9The National Outlook Study (NOS, 2003) states that power production grew 4.8 times from 1980 to 2002, i.e. 6.99% a year on average (fig. 3). This growth outruns the overall population increase (1.8% a year) and urban population growth rates (4.5% a year) – as electricity networks only cover cities at this point. But it is important to point out that the network only reached large cities in the 1980s, but has since stretched to encompass a growing number of medium-sized and even small towns (26 towns had electricity in 2002). In other words, the number of consumers is growing faster than the urban population. For example, 66,000 people were using network electricity in 1995 and 102,000 were doing so in 2001. The annual growth rate here, 7.25%, is similar to the increase in power generation.

10According to this same study, Mali’s urban population should continue to increase at 4.5% a year to 9.15 million people in 2025 (up from 3.2 million in 2002). Plans to hook up every city with more than 20,000 inhabitants and a number of cities counting fewer (93 cities in total) should put the total covered population at 8.4 million inhabitants (44% of the country’s total). As a point of comparison, there were 1.75 million inhabitants (i.e. 16.5% of the total population) in the 26 connected cities in 2002. So it makes sense to say that power consumption will grow at the same pace as it has been growing over the past two decades (6.5% to 7% a year) over the two decades to come. This growth rate suggests that effective power consumption will swell fourfold to fivefold, from 0.43 terawatt-hour in 2003 to about 2.00 terawatt-hour in 2025.

Source: NOS, 2003
Figure 3. Increase in Mali’s power generation

Climatic forcing factors in the Niger river basin

The climate and its several forms of variability

11Hulme (2001) argued that there was no such thing as “normal” average rainfall for the northern Saudi and Sahel areas. What is “normal” here, however, are ample weather swings occurring over several overlapping periods of time. These overlapping variations, however, obscured the forcing factors behind the weather changes and therefore hampered efforts to identify trends and model changes. The term “trend”, as an aside, warrants a word of caution: what may look like a trend over a given period of time may, in fact, be part of a longer cycle.

12Ward (1998) showed the point of distinguishing climatic oscillations spanning different periods of time (or frequencies). We will be following this approach, dissecting the Sudan-Sahel climate’s several components and showing which ones intertwine into what its people can perceive as a “trend” on a ten-year basis.

Variability of the internal structure of the rainy season

13Le Barbé and Lebel (1997) found a link between the Sahel’s dwindling rainfall over recent decades and a drop in rainfall in the middle of the rainy season (in July and August). The rainy season as such, they argue, has not grown considerably longer or shorter. D’Amato and Lebel (1998) go a step further linking this to a decline in the frequency of cyclone passages during those two mid-season months. Thus, the study of changes in the structure of the rainy season can cast light on the mechanisms underlying monsoon processes and, taken a step further, hint at the factors governing the medium-term and long-term trends in total rainfall figures.

Inter-annual variability of total rainfall

14This variability characterises the variations recorded between one year and the next, or between one two-year or three-year period and the next. This variability scale exists but, relatively speaking, is rather less significant than in other parts of Africa. Five-, three- or two-yearly fluctuations only account for 15% to 20% of the total variability. This, generally speaking, means that rainfall figures for one year are roughly like (or at least resemble rather than differ from) those of the previous year. Meteorologists have found this pattern in statistics spanning the last century. But the past decade, conversely, has brought sharp year-on-year oscillations including dry years (1996, 1997, 2000, 2002 and 2004) intertwined with rainy years (1994, 1995, 1998, 1999, 2001, 2003, 2005).

Decade-on-decade variability of total rainfall

15Much of the Sudan-Sahel climate’s variability is due to these patterns (low-frequency fluctuation). Moron (1997) suggested that quasi-decade-on-decade cycles (spanning 12 to 13 years) would explain a full 12% to 14% of the variability. “Typical” dry or rainy periods, in this light, should last about 5 to 10 years. The two dry decades in a row (1973 to 1993) could fit into this variability pattern even if they were abnormally long.

Multi-decade variability of total rainfall

16This category can encompass changes spanning the last century or longer historical periods. There appears to be – certainly since the late 19th century and probably since the 17th century (Hulme, 2001) – a gradual downward slope in rainfall. This downward trend also seems to be affecting the Sahara (which, as an aside, was greener and had more wildlife in prehistoric days than it has today). However, sharp decade-on-decade variability within this historical trend brings very rainy periods spanning a number of years. The latest ones date back to the middle of the 20th century. The future may bring more such episodes.

17Surface conditions seem to amplify the climate fluctuations outlined above, adding a form of inertia to existing rainfall patterns (positive feedback). This could explain why decade-on-decade variability predominates over year-on-year variability.

Changes that people have noticed and noted

18Colonial records dating back to the early 19th century mention a series of damp and rainy episodes. We know that the first half of the 19th century also saw a very dry patch (probably as dry and as long as the 1973-1993 drought). The early 20th century brought another very dry period. The 1950s and 1960s, conversely, were very rainy and brought some severe flooding (fig. 4).

19Today, the notion that the rainfall decline is here to stay is all-too-often considered an indisputable fact and pegged to the notion of desertification. However, desertification is more a case of soil erosion than of the desert sprawl.

Figure 4. The Niger River’s maximum flow rates in Koulikoro
The flow peaks in Koulikoro illustrate the different
variability scales in the Sudan-Sahel (year-on-year
and decade-on-decade patterns, mainly), even if a
dam was built in Selingue in 1982.

20Global warming – another noteworthy parameter, naturally – seems to be hitting West Africa in pretty much the same way as it has been affecting the rest of the world since the early 20th century. How it might reshape the rainy seasons is not yet clear.

The forecasts and outlook

21As opposed to other parts of the world (the south of Africa, for instance), the latest climate-change models from research on global change do not provide stable, unequivocal forecasts for the Sudan-Sahel region through 2100. So it is fair to say that it is impossible to establish long-term climate forecasts in this region. There are, however, models containing scenarios for the coming century as concerns ground temperature, providing very wide rainfall-forecast brackets (- 20% to + 40%). So scientists only agree on the fact that temperatures will rise and that rainfall is unpredictable. We will therefore have to look at more than one long-term scenario.

22However, medium-term rainfall forecasts through 2040 provide fairly consistent results – and provide reasons to hope that the normal figures and indeed slight excess which has predominated since 1994 might stretch on until then (fig. 5).

Figure 5. Observed and simulated rainwater anomalies (variation %)
Source: McCarthy
et al., 2001


Large hydraulic schemes

23Compared to other large rivers around the world, it is fair to say that the Niger is still fairly untapped. This is especially true in the case of its upper and middle reaches. There are only three dams between the springs and Niamey: Selingue, Sotuba and Markala (in that order, downstream). They are all in Mali but vary somewhat in size.


24Selingue is on the Sankarani, one of the Niger River’s tributaries, about 60 km from the junction, 150 km upstream of Bamako. It has regulated the Niger River’s flow to some extent since 1982.

25It is an earth dam. Its two sections measure 600 m in total and nestle the eight 13 x 5 m2 overflow flaps that make up the weir and the central 14 x 11 m2 valve. This weir is designed to discharge a one-in-a-thousand-year flood, estimated at 3,500 m3/s. Four 11.9-MW Kaplan turbines in the power station generate electricity. An irrigation valve diverts water to the 1,500 ha for people displaced by the dam (Selingue, however, only deflects very small quantities of water for irrigation).

26The impoundment area spans two river valleys (65 km long each). When the impoundment lake is full (i.e. when it reaches the 348.5-metre mark in late September every year), it holds 2.63 billion m3 of water and covers 430 km2.

27Selingue dam remains full for a few weeks between late July and 20 September, thanks to the natural annual flood. Then it starts releasing water and feeding its turbines to generate electricity and regulate the flow downstream. This lasts from January to June. The impoundment lake is practically empty (below 10% of full capacity) in early July. So, from an annual perspective, this dam’s impact on downstream flow is simple to describe: it regulates water levels, smoothing swells and compensating lulls. Selingue is indeed the only dam that can do the latter, supplying water to the Office du Niger scheme at the end of the dry season and supplying Niamey (which is nonetheless 1,500 km downstream).

28Administrators are intent on this dam’s lull-counterbalancing role. In particular, they have specified the need to keep the minimum level at 60 cm in Koulikoro and to guarantee a 40 m3/s flow downstream from Markala.

29How filling the impoundment affects the Sankarani River’s flow (and hence the Niger’s flow downstream), conversely, has received little attention. Estimates suggest that Selingue diverts about 400 m3/s from the flow peak (which would otherwise feed into the Niger). Filling about 2 billion m3 in eight weeks involves diverting about 413 m3/s on average during that period of time. It is also worth pointing out that this filling (diverting) phase lasts throughout the Sankarani’s natural swell (peak included), as it stretches on to 20 September.


30Sotuba is on the Niger River, near Bamako. This run-of-river dam produces electricity for Mali’s capital, and diverts a scant 0.22 km3 a year to irrigate Baguineda, a small farming area (which has a negligible impact on river’s flow downstream).


31This run-of-river dam stands some 250 km down the Niger from Bamako, and has been operating since 1947. Its 1,813-metre-long wall and 816-metre-long dam housing multiple-position collapsible spillway gates hold a lake during low-water periods and disappears completely during medium- and high-water periods. Its diversion valve feeds a lake that irrigates nearby Office du Niger schemes by gravity.

32The channels emerging left of the pond comprise the feeder channel (designed to carry 200 m3/s) which splits into the Sahel canal, Massina canal and Costes-Ongoïba canal. These canals feed 77,200 ha of Office du Niger full-water-control land growing rice and market produce, 3,000 ha of controlled-submersion Office Riz Ségou land growing rice and 5 000 ha growing sugar cane in the Sukala agro-industrial area.

33Markala diverts about 2.6 billion m3 a year into Office du Niger schemes. In absolute terms, this is not much (the Niger carries about 46 billion m3 of water a year through this area). But it is about 50% to 80% of the river’s water during the low-water stage. Markala also diverts a fairly substantial 150 m3/s during the swell (returning practically none of it to the river).

34The top-priority goal for agencies administrating river resources is to keep the lake’s level at 300.10 m during the low-water stage and at 300.54 m during the high-water stage, in order to feed the Office du Niger scheme’s areas by gravity. This dam also has to keep the downstream flow at or above given levels during the low-water stage, but can only do so with Selingue help.

35In practice, the Markala dam’s flow is governed upstream. Administrators process information about incoming flow, on-site diversion and outgoing flow for users downstream.

Existing dams and how they impact water flow

36The total amount of water that each of these three dams diverts in a year (table 3) provides the first indication of their impact.

37However, these annual diversion figures do not show how these dams alter flows downstream. Selingue, for instance, diverts very little water in absolute terms, but has a sizeable impact on the river’s flow downstream insofar as it deflects a large amount of water at given point in time, holds it for a few months, and then releases it. So a more comprehensive table is warranted (table 4).

  • 1 Maximum high-water flow rates in Ké-Massina (downstream from Markala, at the entrance of the delta) (...)

38To sum up, the fact that the first two dams (Selingue and Markala) divert water during the high-water stage explains some (500 to 550 m3/s) of the downstream lull. This diversion has a much smaller impact on the swell’s “tail” (late September up the Niger and late October in Mopti) as Selingue is already full. In any case, the negative effect on high-water levels downstream (in Mopti, for example) and on the extent and duration of the Niger’s inland-delta floods is by no means negligible1even if it has never been precisely quantified (doing so, as an aside, would be difficult, as the water stretches across the delta and intermingles with water from the Bani which is not affected, as yet, by the dams).

Table 3 – Diversions by Dams on the Niger River’s upper and middle tributary basin

Table 3 – Diversions by Dams on the Niger River’s upper and middle tributary basin

Table 4 – The impact of these two large dams on downstream flow rates during high-water and low-water stages

Table 4 – The impact of these two large dams on downstream flow rates during high-water and low-water stages

39During the low-water stage (February-June), the backup release from Selingue does a little more than compensate the water that Markala diverts at that time. There is no doubt that flow rates rise in every area downstream from Selingue, but the areas between Selingue and Markala (Bamako, Koulikoro, Segu, etc.) enjoy the bulk of the benefits. Diversion at Markala has a very severe impact on the flow beyond it.



40Plans here involve building a dam with a micro power station between Timbuktu and Gao to generate electricity and to develop farming around the Niger’s oxbow. The lake should hold 6 billion m3, making Taoussa Mali’s biggest project. It will stand on PK1 436, its bed will be 249.95 m deep, and the lake will fill up to 258.75 m. The impoundment would be about 140 km long and stretch nearly as far as Korioumé (Timbuktu’s port) upstream.

41This dam is expected to reshape the annual swelling from the rainy season higher up the basin. When it gets this far north (in December), the swell has already been through the delta and lost some of its strength. It will be even weaker downstream, as some of the water will feed into the lake. So the annual cold-season floods in Gao (January-February) would probably dwindle and disappear. However, this dam would guarantee a 100 m3/s flow rate in Niamey through the dry season, and should irrigate 70,000 to 80,000 ha in Mali.

42Furthermore, a simulation using the Niger River’s mathematical model (Carima) shows that this dam will slow the flow upstream by levelling the slope (once the lake fills to certain level). This, in essence, would lengthen the high-water season – at least north of the delta and downstream from Lake Debo.

43Whether authorities will allow this lake to empty almost completely once a year (as Selingue) or whether they will choose to keep water levels more or less high is not yet clear. The latter would keep the delta flooded for longer.


44This plan involves building a threshold dam on the Bani River, by Talo, to irrigate some 24,000 ha and thereby develop farming. There are no plans for power generation. In Talo, the Bani splits into two canals creating an island. A spillway will close the shallower (north-side) canal (the minimum bed level is 270.50 m) and an embankment will close the deeper (southside) canal (minimum bed level is beneath 268.00 m).

45Simulations show that water would spill at the 277-metre mark. Upstream, water would spill over the left and right banks in two places, flooding a total of 150 km2. Precisely what areas would flood, however, was not clear (the topographical information used for this simulation lacked accuracy). Downstream effects were run through a simulator as well: building a dam in Talo would delay the Bani’s downstream swell by a week, especially in Djenné but should not have a measurable impact on the peaks in Djenné, Mopti or across the delta. This comes as no surprise (this dam is a spillway and fairly small). It should, however, slightly compensate the lull in March and April.


46This project can be regarded as Talo’s rival. The goal, here, is to flood 85,000 ha of plains (the Pondori, which the river’s swell does not reach during poor years) and to generate enough electricity for the neighbouring city. This dam should be small, not unlike Talo, and, similarly, have little if any impact on the rest of the inland delta.

47There are plans for another small run-of-water dam in Kénié, between Bamako and Koulikoro. This dam should have no effect on downstream flow rates.

48Fomi, lastly, is in the upper basin in Guinea. When it is built, this dam will most probably be used to generate electricity (like Selingue). The impoundment lake will probably hold 6 billion m3 of water, meaning it will probably have a substantial impact on the river’s flow rate, and play an even more important role smoothening high-water and low-water flow variations (which, no doubt, will have palpable consequences on flooding in the Niger’s inland delta).

Hydroelectric power generation

49According to a 2004 IEPF study, annual production in 2002-2003 stood at 0.46 terawatt-hour. More than half of that production (56% or 0.262 TWh) came from Selingue (0.18 TWh) and Sotuba (0.04 TWh), on the Niger River. Manantali, a new dam that was starting up in the Senegal River basin at that time, provided backup. When it reaches full capacity, Manantali will be generating 0.41 TWh (at least for Mali; the rest of will go to Senegal and Mauritania). Combined, in other words, these three hydroelectric dams should be generating 0.63 TWh in the near future. Thermal power plants, on their part, were producing 0.20 TWh in 2002-2003 but fresh investment should enable them to generate 0.30 TWh in the near future. Mali’s present capacity, in other words, stands at (or beyond) 0.9 TWh. This covers its needs today.

50But demand is growing fast and capacity could fall behind in a few years’ time – as early as 2007, some say (cf. “How Mali’s society and its requirements are evolving” p. 177). New investment – to at least double capacity by 2025 – will therefore be necessary.

51Hydroelectric power plants will certainly cover some of this capacity. Thermal and renewable sources of energy will also play a role. But exactly how much each source will contribute is not clear yet.


General productivity

52Cotton is Mali’s second foreign-currency earner after gold (612,000 t a year in 2003/2004). Much of it grows in the Niger River’s basin, south of a line between Bamako, Segu and San. This area is referred to as CMDT and spans the Bani, Baoulé and Sankarani tributary basins. Farming, to some extent, can be described as intensive, and fertilisers and pesticides are commonplace. In Mali, cotton is a rain-fed crop.

53Mali’s cereal production has risen from 1,032,000 t a year (the 19611966 average) to about 2,707,000 t a year (the 1998-2003 average). Production stagnated for a time, but has been growing a healthy 5% a year – comfortably ahead of the population since the 1980s. Outside bad-weather years (such as 2004), Mali can typically cover more than 100% of its food requirements.

Figure 6. How Mali’s cereal production and population have evolved since 1961
In spite of sharp year-on-year fluctuation, production has, by and large, grown faster than requirement.

54Cereal production originally increased in step with the amount of land used to farm it, which has stretched from 1,384,000 ha (the 1961-1966 average) to 2,636,000 ha (the 1998-2003 average). Early attempts at intensive farming subsequently increased yields somewhat. Corn yields rose 49% and rice yields 100%. Millet and sorghum yields, conversely, did not increase markedly.

55Rice production was subsidiary until the early 1980s, but has since grown to rank alongside millet as Mali’s leading crop. Most of that rice is grown by the Niger or tributary rivers. A growing portion of the food grown in Mali, it arguably follows, comes from the Niger River.

Table 5 – Cereal production, farmland area and yields in Mali (1961-2003)

Table 5 – Cereal production, farmland area and yields in Mali (1961-2003)


56Rice has become more prominent for two reasons: because rice-farming areas has been growing regularly since the early 1980s, and because average yields have grown concomitantly (from 1 t to 2 t/ha). Rice production idled below 200,000 t a year for a fairly long time, but now invariably reaches at least 700,000 t a year (and peaked at a record-breaking 900,000 t in 2001).

Table 6 – How Mali’s rice production has evolved since 1961

Table 6 – How Mali’s rice production has evolved since 1961

57Rice consumption is higher in cities than in the country, and will plausibly grow in years to come – all the more so as yields still offer considerable room for improvement (the most productive systems produce six tonnes or more per hectare, the least productive ones less than one tonne per hectare).

58Mali counts three distinct rice-growing systems. Their relative share in total output and the amount of land they use vary considerably. These systems follow:

  • river-fed and rain-fed systems are specific to the river valley and, especially, to the inland delta;

  • controlled submersion;

  • full water control (the only ones that can be described as irrigated). Yield is best analysed separately.

Traditional river-fed and rain-fed rice growing (unsupported)

59This ancient system does not involve any hydraulic infrastructure. It involves using local varieties of “floating rice” that can handle water levels of up to three metres. Farmers choose a location for their paddy according to where they believe the water levels will rise during the high-water season (based on their experience). They may also move their paddies from one year to the next. They plough in February or March (after the harvest), or in May or June (just before the rain returns). They scatter the seeds, rainfall triggers the growth process, and then the floods take over. Floating-rice varieties derived from Oryza Glaberrima can survive in three-metre-deep water, provided the level does not rise more than five centimetres a day. Farmers then harvest the rice in the water between late October and December (depending on the variety they farm). This rice-growing method is very common in the inland delta. It spanned about 160,000 ha in the late 1980s (according to the 1988 PIRL project report), and probably spans between 180,000 and 200,000 ha today. Yields are very poor (less than 900 kg/ha on average) and have not changed since the 1950’s. They are also erratic: flooding can be inadequate, and rainfall insufficient or untimely (late rainfall will not allow the rice to grow enough before the flood). Plantation mobility is another of this system’s hallmarks.

60During good years, as many as 10,000 t of rice can be sold outside the delta (Kuper and Maiga, 2002). Bad years bring rice shortages.

61We can look at this system in terms of “usable farmland” (how much land has enough water for rice to grow properly). Table 7 shows that usable farmland can indeed vary considerably from one year to the next. In 1994, for instance, three-quarters of the delta were suitable for floating-rice plantations. In 1972, conversely, only a fraction was viable (Marie, 2000).

Table 7 – Usable farmland, traditional rice plantations, 1972, 1978 and 1994

Table 7 – Usable farmland, traditional rice plantations, 1972, 1978 and 1994

62These sharp variations explain why paddies move about so much. Yield, however, is still unpredictable. Paddies would shift about 10 km in the 1950s (Gallais, 1967), but started moving much further in the 1980s, when flooding became much more erratic. Farmers have left the Niger/Diaka plains and flocked to lower areas (Kootya, in the middle of the delta, for instance). Healthy flooding in 1994, conversely, covered the Kootya area’s hamlets attracting farmers back and triggering new rice-paddy migration.

Irrigated farms (scheme-supported)

63Here, farmers can count on irrigation systems diverting surface water into their farms. A wide variety of systems, ranging from controlled submersion to full water control, fall into this category. The Niger’s tributary basin in Mali counts about 250,000 ha of irrigable land, including some 100,000 ha with full-water-control schemes.

64Mali has ranked water control high on its list of strategic priorities, both to meet its population’s needs and in the interest of sustainable development. Accordingly, it has been working on an irrigation-development plan to tap existing potential (MDRE, 1999). Doing so will involve building hydro-agricultural infrastructure to flood an additional 50,000 ha of full-water-control land and 14,000 ha of land on shallow-flooding, submersion and other systems, by 2007. The longer-term goal is to extend the Office du Niger’s total irrigated area by 120,000 ha, to 200,000 ha, in the next 20 years.

65Schemes using river water for irrigation fall into one of the following categories.

Controlled submersion

66Accounts for 60% of the surface prepared for rice plantations. The schemes are publicly-funded and run by relatively autonomous government agencies. The two most important ones are:

  • Office Riz Mopti (ORM): 34,000 ha;

  • Office Riz Ségou (ORS): 34,676 ha.

67These schemes involve building small walls and inlets at specific levels to govern submersion. This in effect, curbs some of the risk associated with rainfall: in theory, farmers can plant their rice later on in the year and make sure the seedlings are out of the ground before the flooding begins (and authorities controlling the dam can delay the flooding if and as required). But most of these systems were designed in the late 1960s, when the Niger brought healthy and relatively regular swells (the average maximum level in Mopti between 1960 and 1969 was 693 cm). Today, however, these dams are too high up and therefore ineffectual when flooding is inadequate – which, alas, is often the case today (the average maximum level in Mopti between 1980 and 1989 was 539 cm). Average yields are very poor (often less than one tonne per hectare) and only a fraction of sown areas are actually harvested. In ORM plots, for instance, the ratio between harvested and sown areas stands below 0.33 when high-water levels fail to reach the 550 cm mark, and does not reach 0.75 when high-water levels rise above 620 cm. During very bad years (1984), there can be no harvest at all.

68This system is inadequate due to swell variability and uncertainty. In most cases, it can not guarantee production or earn farmers a decent income. It is inflexible, as well: rain- and river-fed systems may be less productive but can, at least, move to where conditions seem most propitious.

69The ponds have been arranged slightly differently around Timbuktu and Gao (the Action Riz Sorgho project). The goal, there, was to build small structures to slow receding water and thereby keep the paddies flooded for longer.

Gravity-governed full water control

70These schemes use upstream impoundment lakes with guaranteed water which release water by gravity. The Office du Niger’s “large-scale” schemes (spanning 83,000 ha, including 5,000 for sugar cane), Baguineda (3,000 ha) and Selingue (1,000 ha) are three examples. They are all linked to dams (cf. “Large hydraulic schemes” p. 184).

71Water management, here, involves building a network of canals (of earth, usually) to route water from the lake to the farms, crisscross it around the plots, and then allow it to flow downstream into the drainage network.

72This rice-growing system ranks third in terms of farmland size (it covers slightly less than 100,000 ha across Mali). But it produces nearly 80% of this country’s rice in an average year (different authors put yields between 4.5 and 6.0 t/ha). These areas have also attracted sizeable populations (farmers and their families, mostly). The extended Office du Niger area (i.e. all the land under Office du Niger management) spans 28,174 km2. Its eight hydraulic schemes span 19,074 km2 and were home to some 321,069 people in 1998, putting their average population density at 16.83 inhabitants per km2 (Sogreah, BCEOM-Betico, 1999).

73These schemes, in other words, are efficient. But running them can entail a number of problems and they have a number of sensitive aspects. Monitoring and servicing heavy-duty hydraulic equipment is one issue. Administrating the land and cost distribution (water costs, in particular) among the parties is another sensitive area. Protecting the soil’s fertility and, more generally, preserving the environment, is yet another concern. Public agencies in charge of these large areas are working hard to fine-tune and implement sustainable policies in line with central-government guidelines. Doing so, however, entails hiring qualified staff and experts – which, of course, carry a cost as well.

74Another issue, from an ecological standpoint, is that most of the water that this large system diverts from the river is never replaced.

Table 8 – The Niger River’s supply at Koulikoro and water diversion in gravity-governed total-water-management systems

Table 8 – The Niger River’s supply at Koulikoro and water diversion in gravity-governed total-water-management systems

Source: Kuper et al., 2002

75These large systems, in other words, divert considerable amounts of water (table 8), and therefore have a negative impact on downstream water availability during both seasons: during low-water stages they divert more than 50% of the river’s waters (and as much as 80% at peak times), during high-water stages they divert water into gravity-governed schemes and into Selingue’s impoundment lake concurrently.

76How much water is diverted into large gravity-governed systems (the Office du Niger’s schemes, mainly) of course hinges on the size of the area in question. But it also hinges on the system’s efficiency (i.e. its ability to divert no more water than the irrigated areas need). These irrigation systems, however, appear to waste water at present (cf. “Poor scheme water efficiency” p. 253).

77We can also measure efficiency in terms of how many kilograms of rice are produced with a given amount of diverted water. In the case of the Office du Niger, one cubic metre of water produces about 0.13 kg of rice

78(320.106 kg divided by 2,500.106 m3). This, of course, is after deducting the water that goes to sugar-cane plantations and non-farming uses (drinking water for livestock, etc.). This is also mediocre: the “norm” in Asia, for example, is at least 0.2 kg of rice per cubic metre of water.

Figure 7. Size of Office du Niger areas by scheme type, crop and season

79Poor water efficiency seems to be the weakness affecting gravity-governed schemes such as those run by the Office du Niger. It is also true that scant pressure on water resources in recent years did little to promote efficiency.

80Higher efficiency, however, is one of the key issues in efforts to develop these schemes and to improve overall river-water management. It is worth pointing out that large gravity-irrigated schemes are expanding at a steady pace and should continue to do so in years to come (fig. 7).

81This chart shows a sharp increase in total surface, and in out-of season plantations (hot-season rice and vegetables).

Pump-assisted total-water-management irrigation

  • 2 This system is also in use in Kayes, in small areas feeding off the Senegal River.

82Pump-assisted irrigation started developing here in the early 1980s, when rainfall and swells started dwindling. The main schemes on the Niger River2 are in Mopti, Gao and Timbuktu. These schemes are generally small, and run by private concerns or village communities, i.e. “Village Irrigation Areas”, often backed by NGO-sponsored “projects”.

83The inland delta counted 154 of these operations, spanning a total 1,400 ha, in 1998 (Ducrot et al., 2002). In these schemes, a motor pump feeds a stilling pool and water then spreads across the plantations by gravity. The results are usually good or very good: yields generally exceed 6 t/ha, and can be as high as 9 t out of season (in the dry season). The problem, here, is that building these schemes entails considerable investment and that they will only prove cost-effective if the area they cover is large enough. Experience, especially in the Senegal valley, has shown that these schemes are not cost-efficient in small operations. Operating costs (high pump-fuel costs, especially), community management, and heavy maintenance requirements can cause a number of problems, and indeed lead users to abandon the entire scheme or at least part of it after using it for a longer or shorter period of time.

Low-water and off-season farming on the river banks

84These schemes are especially developed in the north of the Niger River’s delta and oxbow, and in the lake region. There are seventeen large lakes, mainly on the right bank. The shores of these lakes, and the shores of a large complex of permanent and semi-permanent ponds, are used to grow sorghum or tubers and vegetables (potatoes, tomatoes, etc.) as soon as the swell starts receding in late January or in February. These receding-flood crops are vital for people living in these Sahel or desert areas, who can no longer count on rain-fed crops for subsistence. The size of the tillable area, here, is linked to the area covered by the flood, and therefore linked to swell levels. The potential, however, seems largely untapped.

Table 9 – Different crops feeding off the river, their requirements, and their impact on the flow

Table 9 – Different crops feeding off the river, their requirements, and their impact on the flow

Agricultural pollution in the river

85Intensive farming in Mali may not be as commonplace as in northern countries, but efforts to increase yields nonetheless entail using more fertilisers, pesticides and machinery. Malians are farming more land, as well. Combined, these factors are accelerating soil erosion and allowing more of the soil’s elements to seep into the river. Intensive farming is gaining ground in the cotton plantations on the Niger River’s basin, so pollution from chemical fertilisers and pesticides is likely to follow and intensify. Intensive farming methods are also moving into irrigated areas, and likewise bringing fertilisers and pesticides (but in very small quantities as yet). Growing farming areas, escalating intensive farming and developing agricultural-industrial complexes will heighten the risk of pollution, spelling the need to monitor water pollution now.


86A small portion of Mali’s fish production comes from the Senegal river basin (from the Manantali dam lake). The bulk of it comes from the Niger River basin. As do the production swings, which can range from 55,000 t one year to 140,000 t the next (fresh weight equivalent) (fig. 8). The bulk of the Niger River basin’s production, in turn, is fairly concentrated: about 80% of it (40,000 to 120,000 t a year) comes from the Niger River’s inland delta (broadly, from Ké-Massina to Diré, encompassing the lakes). The rest comes from the Selingue’s impoundment lake (4,000 t a year), from the Niger River upstream of Massina, from the Niger River’s oxbow between Diré and Gao, and from the southern tributaries (Baoulé).

87This healthy production makes Mali practically self-sufficient in fish, at least when it has average or generous high-water stages (which put production in excess of 100,000 t). Small amounts of frozen sea fish are imported and consumed in the capital, and some fresh and smoked fish is exported to Burkina Faso and Côte d’Ivoire.

88It is not easy to tell how many fishers there are in Mali for two reasons. First of all, it is not easy to tell what exactly defines a fisher, and secondly statistical coverage is inadequate. We can nonetheless borrow international FAO terminology to establish what exactly defines a fisher. Some people fish sporadically, but are primarily farmers. They can be referred to as occasional fishers and come from a wide variety of ethnic backgrounds. Other people split their time more or less evenly between fishing and farming. They generally belong to large families. The several family members involved in fishing and farming usually split the work from December-January until May-July: some of the men stay in the village to harvest the fields, rest and then tend the rice paddies; the others leave the village and fish intensively until the rainy season leads them back to the village where they spend the rest of the year working the fields with their relatives. These people are referred to as part-time fishers. Some people and households live exclusively from fishing, and often migrate to wherever fish is most abundant. These people are referred to full-time fishers. In Mali, part-time and full-time fishers are generally bozo or somono (Fay, 1989; Bauman et al., 1994).

Figure 8. Fresh fish landing (in tonnes)
Recent changes in Mali’s fish production as reported by national agencies to FAOSTAT. The bulk of production (over 95%) comes from the Niger River basin.
Source: FAOSTAT (updated in early 2004)

89Large-scale surveys in the late 1980s narrowed the number of “real fishers” (part-time and full-time) down to around 30,000 in the delta alone. This suggests that about 60,000 men fish (Morand et al., 1991). Those figures, no doubt, have grown since. And they do not include fishers from other parts of the Niger River basin elsewhere in Mali. Including them all, there are plausibly at least 80,000 fishers today.

90But there are also people working in other areas of the fishing industry: there are fillet merchants, fish merchants, fish processors and jetty hands (box carriers, packers, etc.).

91So, if we consider all these categories and accept the admittedly minimalist notion that every fisher (strictly speaking) generates one job on land, we can say that fishing provides work for 160,000 people. Then, if we factor the accepted notion that rural West Africa counts two people who are not engaged in professional work for every person who is (Hugon, 1998), we can say that about 500,000 people (children and elderly included) live off fishing (i.e. fishing accounts for all or for a sizeable portion of their livelihood). This estimate may be rough, but it nevertheless points to the fact that fishing plays an important role even if it only contributed a scant 4.2% of Mali’s GDP in 1995 (Breuil et al., 1996).

92Fishing campaigns last eight to nine months, in all this river’s areas. Intensive fishing begins when the floods recede in November or December (depending on the area). They last through the low-water stage and end with the following swell in late June or early August (again, depending on the area). As fishing is intensive, catch sizes decrease as the campaigns move on. Estimates suggest that a full three-quarters of the river’s fish biomass in early December is gone by the end of the campaign (Kodio et al., 2002). Then, however, the high-water stage gives the fish population time to grow back to its original biomass between July and November. Biomass reconstruction, in this case, is complete – at least if the high-water stage is good. The catch one year has never been seen to hamper the catch the next.

93Outside Selingue lake, where it is remarkably stable, fish production across the Niger River valley in general – and in the delta in particular – can vary considerably from one year to the next. Fish production is closely correlated with swell levels and with how long the flooding lasts (fig. 9). This is because the fish that hatch at the beginning of the high-water stage (in July) grow up in the flooded plains through November (Bénech and Dansoko, 1994; Niaré, 1994). They are then caught when the flood starts receding and the next fishing campaign begins.

94Fish production can also be measured in relation to a producing area’s size. In the inland delta, for instance, the flooded area spans 7,000 to 20,000 km2 (depending on the year) and produces 40,000 to 120,000 t of fish (depending on the year and on flooding, Laë, 1992b). So, in this case, one flooded hectare produces 60 kg of fish on average. This is within the generally accepted bracket for this type of environment (40 to 100 kg/ha, Welcomme, 1989). Intensive campaigns in particularly productive spots on the delta, however, can produce as much as 100 kg/ha or more.

Source: Morand and Kodio, 2004, based on Observatoire de la pêche figures
Figure 9. Median catch per unit of effort (per fishing trip, in this case) against a swell index (how many days the water level stays above the 4.50-m mark in Mopti)
The round dots and top line show figures for the first half of the fishing campaign (December to mid-March); the squares and the bottom line show figures for the rest of the campaign (mid-March to June). The numbers denote years.

95Village fish farming still has considerable potential, but in other areas (outside the large-scale commercial fishing areas) and with other population groups (more farming-oriented populations, especially) (Niaré et al., 2000).

96From an environmental perspective, in its current form, fishing has no impact on the rate or amount of water flowing or lying anywhere in the Niger River. It most probably generates very little pollution at this point (only the hydrocarbon released by the engines of vessels carrying migrating fishers or fish).

Table 10 – Water requirements for fish and fishing

Table 10 – Water requirements for fish and fishing

Livestock farming

97Estimates suggest that there are 7.6 million head of cattle in Mali. Figures are relatively stable in the large traditional cattle-breeding area in the north of the country (Mopti), and growing in the south. In Sikasso, for example, there were 336,000 head in 1970 and 741,000 (121% more) in 1995. However, in absolute terms, Mopti is still Mali’s leading livestock-farming area. An aerial census in the 1980s suggests that there are some 1.6 million head, including about 1.2 million in and around the Niger River’s inland delta.

98The inland delta plays a very special in Mali’s livestock-farming sector. This is because its flooded pastures are exceptionally fertile and stay “green” for 7 to 8 months. The best pastures, the bourgoutières, produce between 20,000 and 30,000 kg of dry matter per hectare per year and supply green feed from December to July (Sahel pastures, as a point of comparison, rarely produce over 2 t of dry matter and are only green for four months a year). There are also bourgoutières in the Niger River valley upstream from the delta and (especially), downstream from it, mainly in the oxbow, where they play a central role in the organisation of the Gurma area’s livestock farming. These areas, however, are not very big.

99The inland delta, in other words, is the only natural region in Mali – and in West Africa – that provides generous pastures over vast areas. It spans only 1.5% of the country’s land but counts about 20% of Mali’s cattle.

100When the swell starts flooding the inland delta in August, herders leave on their long seasonal migration to the Sahel’s pastures. They return in November or December, depending on how fast the water recedes. They take their herds across the delta through a complex of trails and shelters, which they use by strict unwritten traditions.

101Other than the water that cattle drink, herders do not, strictly speaking, consume water resources. Yet, not unlike fishers, they depend on the annual flooding (high-water levels dictate how much fodder they will find for their herds).

Table11 – Water requirements for cattle herding

Table11 – Water requirements for cattle herding

Table 12 – Delta fodder production and theoretical change based on different high-water levels in Mopti

Table 12 – Delta fodder production and theoretical change based on different high-water levels in Mopti

Source: Marie, 2002

102Poor swells, in other words, can be a threat to livestock farming (plant biomass renewal hinges to some extent on the flooding).

103Poor flooding (5.1 m) will cut fodder production by almost 40%. The same swell (5.10 m), however, would flood 80% of the area that a 6.6 m swell would flood. So the loss in fodder production is considerably less significant than the drop in flooded areas. This is because plant formations that are not under water can still grow thanks to rain and rising groundwater. So flood variability does not affect fodder production in a straight, directly-proportional way: changes in total flooded area only account for some of the change in fodder production. In environmental terms, we can infer that the delta can support 1.5 million head during a very good year, but only sustain 900,000 during a bad year (after a poor flood). Fodder production, however, seems to be resilient: after one or two poor floods and commensurately poor production, a healthy swell will allow fodder to thrive again (Hiernaux and Diarra, 1986).

Household consumption and waste

104From a quality standpoint, household impact on river water can be summarised as follows (table 13).

105The problem can come from quality requirements – especially as domestic circuits can deteriorate water quality commensurately, at least locally (in and around towns and villages) and at specific points in times (low-water stages, rainy seasons and high-water stages).

106Wastewater can fall into one of two categories: black water (from WCs or latrines) or grey water (soapy water from baths, showers, dish washing, etc.).

107Leroy (1999) estimated that town and village dwellers produced 65 g of dissolved organic matter and 70 g of suspended matter a day.

Table 13 – Water requirements and impacts, household consumption

Table 13 – Water requirements and impacts, household consumption

Industrial consumption and waste

108Urban and industrial waste can vary considerably, and encompasses effluents from slaughterhouses, dyeing plants, factories and hospitals, compounded with refuse from market streets and pavements. Their impact on the Niger River varies as well: slaughterhouse waste (which is organic) is locally contained and less worrying than factory waste.

109The district of Bamako’s factories produce 9,715 metric tonnes of waste a year. Of that total, 7,830 t (81%) are recovered, 276 t (3%) are recycled, 238 t (2%) are burnt and 1,372 t (14%) are discharged (ESPGRN Sotuba). The eight largest industrial plants are believed to release an average of 2,200 m3 of wastewater a day (i.e. about 800,000 m3 a year) and the 300 dyeing plants 16,000 m3 a year. This wastewater seeps through the soil or pours directly into the Niger River.

110A survey by the Direction régionale de l’assainissement, du contrôle des pollutions et des nuisances (DRACPN, a regional agency in charge of sewers, pollution and nuisance control) in the region of Koulikoro showed that several industrial plants are responsible for polluting the river, and that 80% of Mali’s factories are between Bamako and Koulikoro. It also showed that the Niger River’s main source of pollution is the Mali’s cottonseed oil factory (Huicoma).

River transport and pollution

111Two types of vessels carry people and goods along the Niger River: two large ships operated by Comanav (malian sailing compagny) and a host of smaller privately-owned boats, generally weighing 10 to 50 t. Fishers also use 2 to 4 t smacks for their seasonal migrations with their families.

112Generally speaking, the Niger River is not fit for year-round navigation. It is not deep enough during the low-water season. Large ships can only use it from early or mid August through mid December or early January, depending on how long the high-water stage lasts and how much water it brings (both of which generally vary in synch). The stretch between Koulikoro and Mopti is navigable at the beginning of the season. Koulikoro-Timbuktu and Koulikoro-Gao routes open some time after that. As the season draws to an end, ships can only use the downstream stretch (after Mopti). Smaller vessels can use the river until February and family smacks can use it until the end of the low-water season (at the risk of getting stranded).

Table 14 – Water requirements and needs, river sailing

Table 14 – Water requirements and needs, river sailing

113These boats produce a mixture of hardly-quantifiable forms of pollution. Generally speaking, however, it is fair to say that Comanav’s two ships and the many smacks using this river do not release substantial amounts of pollution. Noteworthy pollutants, however, follow:

  • motor oil (with little impact, as people often use spent oil for other purposes);

  • wastewater from lubricant-tank washing;

  • oil leaking from defective vessel engines or tanks.

114Some of the waste from exchanges in vessel-staging areas (plastic bags, oil and grease) can be toxic. Pollution levels hinge on how big the staging area is, and on how long vessels stay there. During a journey, all the waste produced onboard is discharged into the river, along with rotting merchandise. Washing oil and pesticide tanks also pollutes the water. Table 14 summarises this.


Surface water quantities

115Surface water levels hinge more drastically on the weather – and on extreme rainfall-figure swings – than anywhere else in the world, here (cf. “Climatic forcing factors in the Niger river basin” p. 180). The drought in the 1970s and 1980s is greatly responsible, on the one hand, for the sharp seasonal variations we are seeing today (alarming low-water levels) and, on the other, for the sharp year-on-year changes. There is still some debate as to whether flow rates have embarked on a downward trend compared to rainfall (i.e. as to whether the same amount of rain as in previous years produces a weaker flow today).

116Some of the surface water, especially in the north and northeast of the Niger River’s inland delta, is in lakes that only fill when swells rise beyond certain thresholds. In these cases, the connection between rainfall and surface-water resources is less straightforward. These lakes are downhill from the river, and quite a way away from it. They fill up when the river water reaches a certain threshold and spills over into the channels that carry it downhill (Orange et al., 2002). When the swell recedes, the water that spilled into the lakes stays there (it does not pour back into the river). As these lakes are bowl-shaped, they can hold water for as long as two or three years once they fill up. Evaporation takes its toll, these lakes keep the bulk of their surface water. In other words, provided swells reach necessary thresholds once every two or three years, these lakes will provide water resources. Or, otherwise put, the occasional poor flood will not affect them. Conversely, whether a swell is poor or very poor makes no difference, as the water level will not reach the threshold either way (as was the case throughout 1981-93). Historical figures of water levels in these lakes will help us determine the high-water threshold that river waters need to reach in order to fill them.

Ground-water levels

117Mali has generous amounts of underground water. The annual renewal rate stands at 55 billion m3, but only 105 million m3 (a scant 0.2%) are actually used.

118Water reserves fall into one of two large categories, depending on where the water comes from. There are extensive aquifers and more local fractured aquifers.

119The extensive aquifers are associated with poorly-consolidated or nonconsolidated detrital deposits, essentially of continental origin, which gathered in sedimentary basins during the Mesozoic and Cainozoic. They generally comprise several layers, feature inter-granular porosity, and span a little more than half this country’s territory. The ancient waters they enclose are generally not renewed due to the scant rainfall and to the fact that they lie too deep for new water to reach them on a regular basis. More than half the wells cut into these aquifers can provide over 10 m3/h, and some of them can produce more than 50 m3/h.

120Fractured aquifers are associated with crystalline deposits (the base) or Precambrian and Palaeozoic sedimentary deposits. They have continuous or semi-continuous tables, depending on the density of the fracture networks in them. There are discontinuous aquifers, semi-continuous aquifers (which receive considerable amounts of rainfall due to their geographic position), and surface aquifers.

Water quality

Surface water quality

121Research on the quality of the Niger’s waters (Iwaco, 1996; Equanis, 1997; Gihrex, 2000; Ghenis, 2001) suggests that chemical contamination is as yet only slight, not unlike other African rivers (Picouet et al., 2002). This river’s waters are by and large – and generally across Mali – free from metal pollution, nitrate and ammonium pollution, phosphate pollution and biological pollution due to the large amount of water and its dilutive effect. This is logical, as there are few industrial complexes on the Niger River’s tributary basin, farmers do not use large amounts of fertilisers and pesticides, and the area’s population density is low.

122However, as regards this last aspect, weekly monitoring of the quality of well and drilling waters and of surface water in the cotton-growing area (CMDT in south Mali) showed isolated but noteworthy pesticide-pollution spikes (DDT, especially), associated with fertiliser storage issues or unsuitable use (Bonnefoy, 1998).

123Moreover, if the Niger River can by-and-large boast healthy physical and chemical properties – other than in specific points near industrial sewers between Bamako and Koulikoro (Palangié, 1998) – there are isolated bacterial-pollution phenomena in most dwelling areas. There, the high concentration of colon-type bacillus and faecal streptococcus make the surface water unsuitable for drinking by WHO standards. Dwellers, however, drink this water. So there is a considerable health hazard.

  • 3 Biological Oxygen Demand, an indicator of biological pollution.
  • 4 Chemical Oxygen Demand, an indicator of chemical pollution.

124Moreover, the stretch between Bamako and Koulikoro warrants special attention given the high concentration of housing as well as industrial and farming operations its banks harbour. Work monitoring eighteen risk points (Palangié, 1998) on that stretch showed that water in every effluent-inflow area was unfit by WHO standards on at least three scores. The top-priority issues are BOD3, COD4, suspended matter and heavy metals. However, in every case, BOD, COD and suspended-matter content is at the same level or below that of wastewater in first-world countries, with the exception of wastewater from the Dar-Salam power plant (which is, by far, the most polluted). Our measurements moreover show that a considerable amount of the polluting load is linked to suspended matter measuring more than 0.7 µm. So simply decanting Bamako’s wastewater in suitable facilities (yet to be built) should allow authorities to eliminate a substantial amount of pollutants.

125To conclude, all surveys point to the fact that the quality of the river’s surface water is still generally good outside specific areas or stretches. But it is under threat: large-scale industrial projects, agricultural development in the Niger River valley, and the fact that riverside towns systematically discharge untreated effluents, suggest that pollution may well become a serious issue in the future.

Ground-water quality

126There are two problems with ground-water quality, and they are both in the district of Bamako: one is Bamako’s surface water table, and the other involves the deep aquifers beneath it (Palangié, 1998).

127Bamako’s surface water table is completely polluted: it is very near the surface, hence near the many individual sources of pollution in that district. There are practically no pollution-curbing measures in place and the soil is highly permeable in certain areas (on the right bank, especially). There, many unprotected wells and poorly-designed private sanitation facilities (Zallé and Maiga, 2002) open up direct links between the surface and underground waters. Several studies (Alpha et al., 1991; Iwaco, 1996; Palangié, 1998) have characterised the pollution and found a high concentration of nitrates, nitrites, phosphates and chlorides, possibly from a variety of sources (detergents, fertilisers or decomposing organic matter). The water may also contain pesticides, and certainly contains very high amounts of bacterial pollution (colon-type bacillus and faecal streptococcus, total germs) from diffusing wells that are either too deep or too small, from pervious pit toilets, or from faecal matter carried by surface water through wells (which generally lack proper rims). The water in this surface table, in other words, is unsuitable for human consumption by WHO standards. The rampant pollution that has been recorded in recent years, compounded with the mushrooming population and Bamako’s urban sprawl, will soon deprive most of the population of its main water resource (66% in 1998) if nothing is done.

128Bamako’s deep aquifer (fractured sandstone aquifer) is all the more worrying as it is regarded as a source of drinking water for the near future. This aquifer connects to the surface table through fractures and major accidents which open as far as 100 m deep. Thus far, analyses have shown that the quality of the water in this fractured aquifer is very good. But drilling into the deep water table is likely to change that. Opening up tunnels to the surface will not only open the door to the pollution lying there (if suitable measures to protect the system are not implemented). Most particularly, it will ease the pressure that has thus far stopped some of the surface water from seeping into the deeper table. If pollution levels in the surface table worsen, serious pollution in the fractured sandstone aquifer will become a real possibility.

129Pollution, it seems, is not a major issue in Mali’s other surface and deep aquifers, given the overall quality of its surface water. Only pinpointed areas such as water tables under large cities, can provide the odd exception.

Geomorphological stability: erosion/sedimentation/sanding


130The rainfall, landscape, basin lithology and resulting soil quality explain the bulk of the Niger River’s erosion rates. Studies into the geographic variation of suspended-matter flows by DNH with Orstom then IRD from the 1970s until the 1990s show that erosion is higher on the stretch between Koulikoro and Ke Massina due to changes in lithological conditions (Picouet et al., 2002). Exportation of matter from the Niger River’s upstream basin, as an aside, is remarkably lower than other large basins around the world. It is also worth pointing out that, towards the desert, soil vulnerability to wind erosion increases substantially and that flow of matter (sand in that case) increases in the river after the inland delta. This gradient suggests that human or farming activities which are much more significant in the south (upstream from the basin) do not generate considerable erosion as yet.


131Given the low concentration of solid matter in suspension in the flow, there is little overall sedimentation in the Niger River valley. However, research has shown a certain sedimentation phenomenon in the Niger’s inland delta. The loss of total suspended matter between the stretches entering this area (Ke Massina and Douna) and the stretch emerging from it (Diré) shows that this delta retains between 0.15 million t of sediments (during “dry” years, when the flood covers a comparatively small area) and 1.3 million tonnes of sediments (during heavy flooding). In fact, the loss of sediment across the delta is practically directly proportional to the suspended matter that enters it. But a closer look shows a difference in sedimentation patterns in the upstream delta and downstream delta: sediments invariably build up in the upstream delta, but the downstream delta’s inflow and outflow are relatively balanced.

132The upstream delta is a vast, classic flooding plain. Its numerous meanders feeding countless ponds and generous vegetation provide perfect sediment traps. In this part of the delta, the amount of matter that settles hinges directly on the swell’s flow coefficient. However, the actual amount of sediment settling here is very low compared to the amount on the surface of the delta. There is no risk of the delta filling up, as it were. The sediment, to the contrary, contributes generously to this area’s fertility.

133By the time the water reaches the downstream delta, it has been through the upstream plain and central lakes. Its sediment content is therefore scant and the amount of sediment that settles practically negligible. The remaining swell in the smaller beds and probable versant erosion at the beginning of the high-water stage, then the remobilisation when the flood starts receding paired with wind erosion after the high-water stage at the beginning of the dry season are no doubt the key processes shaping this downstream delta’s sedimentation patterns.

134Lastly, Gihrex research showed a specific threat on the Debo, a large lake in the centre of the inland delta. Decades of droughts have allowed the bourgou to thrive on the banks (farmers even planted it). This fodder plant traps terrigenous sediments brought by the river, and blends them into organic matter year after year. However, despite the fact that flow rates have been on an upswing for the last ten years or so, a silt plug at this lake’s exit has formed. This plug could jeopardise the lake’s ecology, navigation and, eventually, the Niger River’s life (Orange et al., 2002).

River “sand silting”

135Sand silting is a common term that can encompass different phenomena (sand-stock growth, sand-bank shifts, and factors such as receding beaches and bank destabilisation which can contract the river bed in given areas).

136Scientific findings concur that the Niger River bed’s sand stock is not swelling in Mali. Sand banks, however, are shifting. This is an issue for river users and stems from the fact that no fairways are dredged (fairways are otherwise systematically dredged in rivers with sand beds).

137Beach depletion and some bank instability stem from exploitation of riverbed sand near urban areas. Sandbanks appearing (so to speak) when water levels drop can give the impression that the river is sanding up, especially during particularly dry spells (June 1999, for instance).

138Lastly, in the Niger River’s oxbow, wind-borne sand wipes out crop areas and locally clogs channels and other watercourses feeding into ponds or basins. This chokes meanders in the river’s secondary branches: the Koli-Koli, for instance, has become unfit for navigation.

139The reverse, sand depletion, is also a noteworthy phenomenon – and caused exclusively by human activity. As the extracted sand is used for building, this phenomenon is predictably acute around large cities, especially Bamako and Koulikoro. This stretch has seen its sandbanks recede about 100 km in 10 years! This is altering the river’s ecosystems: river plants have disappeared taking spawning beds with them. It is also altering flow patterns and increasing the risk of flash flooding. Sand extraction is the key environmental issue for this river.

Soil fertility

140In Mali, on the whole, declining soil fertility is one of the single most important factors inhibiting crop production. Generally speaking, demographic pressure (entailing shorter fallow spells), as well as water and wind erosion, are the causes. However, erosion phenomena, when they occur, often only shift soil fertility without much consequence otherwise.

141In the Niger River’s inland delta in particular, an in situ combination of the ecosystem’s natural processes and of related traditional production systems (livestock farming in particular) support soil fertility. However, the sediments that the river brings from its upstream basin (in moderate, yet not negligible, quantities) supplement this fertility (all the more so as its generous swells bring considerable flooding).

142In large irrigated areas (such as Office du Niger areas), soil fertility is not only affected by an increase in alkalinity, but also by a drop in organic and mineral fertility (potassium), which is starting to have a perceptible impact, especially in the area of Massina. Soil salinisation is also an issue in these areas, but an in-depth understanding of this phenomenon and its associated risk factors enables authorities to contain it more efficiently. Some risk factors, however, are intensifying. These follow:

  • extending crop areas and shrinking water availability curb leaching;

  • ongoing irrigation-canal use to feed off-season crops helps refill and raise water tables, which in turn promotes salt concentration on the surface;

  • a proliferation of cursory schemes (poor sloping and drainage).

Plant communities and resources


143Observers agree that Mali’s forests are in danger. How fast these forests are shrinking, however, is unclear. Woody vegetation, it seems, has been shrivelling a fairly limited 0.05% a year over the past fifteen years (87 km2 a year, 1,000 km2 in total), essentially making space for farms and due to the fact that several forests in the region of Mopti have disappeared. Deforestation, in other words, seems contained – contrary to what villagers feel. This may be because the natural forest reformation is doing its work.

144By and large, the drought has not led species to extinction or seriously affected biodiversity in Mali’s flora. Some species, however, are showing signs of depletion.

Plant communities in the Niger River’s inland delta

145ILCA scientists found 189 species (137 herbaceous species and 52 woody ones) in the delta. This is relatively poor compared to the 350 species in neighbouring Gurma – but akin to most flooding plains. This poor flora, however, contrasts with vigorous seasonal biomass production, which is 15 to 20 times higher than in the neighbouring Sahel areas.

The delta’s forests

146According to the Wood Resource Inventory Project (or Pirl, for Projet inventaire des ressources ligneuses), forest stands cover 4.5 million ha across Mopti, and plantations and fallow land span 1.8 million ha. Sparse shrub savannas and a variety of palm stands make up the bulk of those areas. Different thicket varieties are part of the landscape as well.

147The inland delta and north-western Farimaké plains count three different stand families. Their sizes differ considerably. The early-1980 figures follow:

  • flooded Acacia kirkii forests withstanding 4-metre-deep water. These forests span 5,000 ha and play a critical role in water-bird reproduction;

  • Acacia nilotica and Acacia seyal forests, mainly watered by percolation, but sometimes secondarily watered by swells (water levels do not exceed 0.6 m). These forests span 115,000 ha;

  • dry woodlands growing on mounds or skirting the delta, spanning 425,000 ha.

148South of this area, we can also find park-type stands made up mainly of Faidherbia albida, a tree that also grows on the sandy savannah benches along with Andropogon gayanus and surrounding rice paddies.

149There have been no surveys into changes in the inland delta’s vegetation in general or into its forests in particular. Observers agree that the forests have been facing serious threats for the last three decades, due to the compounded effects of the drought, scant flooding since the late 1970s, and anthropogenic pressure on resources. Arguably, land-development laws and authority malfunction before 1991 – which replaced the handed-down land-administration rules that local breeder, fisher and farmer communities had applied thereunto – explain a lot of this destruction (Marie, 1989; Moorehead, 1997).

150The Pirl survey suggests that forests in the Mopti region could produce 1.8 million tonnes of energy wood a year (i.e. that forests could export that much wood a year without damaging their “capital”). The Mopti region’s total energy-wood stock, estimates suggest, totals 40 million m3 (including 3.5 million of deadwood).

151Based on Pirl research (1988), the Domestic Energy Strategy (or SED for Stratégie énergie domestique) recalculated the annual production of areas covered by the delta’s various forest stands (table 15).

Table 15 – Wood potential across five circles in the Niger River’s inland delta

Table 15 – Wood potential across five circles in the Niger River’s inland delta

Note *1m3 = 2.3 stere et 1 stere = 330 kg Source: SED project, 1999
Source: SED project, 1999

152There seems to be abundant standing deadwood, especially in the north of the delta. This stock could supply 70% of the firewood needed in the Niger River valley.

The delta’s grasslands

153Most of the inland delta’s plant formations are grasslands, encompassing grassland savannahs and aquatic prairies proving exceptionally rich pastures. Varieties include the following:

  • bourgoutières (floating grass). This grass grows on plains withstanding 4 to 5 m of water a year, but only survive the deep-water submersion if the water levels rise slowly. These are the delta’s most productive courses (20 t/ha) and span about 160 000 ha;

  • vétiveraies cover about 30% of the delta (640,000 ha). Their average production varies from 10 to 12 t/ha, depending on environmental conditions;

  • wild rice paddies cover 15.4% of the delta (about 340,000 ha) and produce between 6 and 8 t/ha;

  • eragrostaies cover 190,000 ha. Their grazing value varies between 5 and 8 t/ha;

  • very low-flooding varieties (andropogoneae, panicaie, etc.) covering about 200,000 ha and producing around 5 t/ha.

154There is not much information about how the inland delta’s grasslands have developed in relation to poor flooding in the 1980s and 1990’s, due to the lack of time-sensitive studies. But research has shown that flooded grasslands in this area reach their ecological balance when flood rates in Mopti average out at 6.60 m.

155Based on this benchmark, we have to distinguish passing developments caused by a temporary decline in flood rates (meaning reversible trends and the assumption that plant formations will regain their balance when healthy floods return) and irreversible environmental developments (which would usher in a new balance between high-water rates and plant formations that swells of under 6.60 m can support).

156Concerning the first hypothesis (that the dwindling flood rates are temporary), a 1986 Hiernaux and Diarra study conducted after the very poor 1984 high-water stage (the century’s lowest) showed the following:

  • deep formations (bourgoutières, deep vétiveraies, etc.) are relatively unaffected by poor swelling. Neither indeed are higher ones (panicaies, andropogonaies…), which depend more on rainfall than on swelling rates;

  • intermediary formations (orizaies, eragrostaies, high vétiveraies, etc.), conversely, can see their production and structure severely affected. The dominant species is decimated and replaced by very irregular populations, dominated by Sahel annual varieties.

157So it seems that this delta’s deep formations as well as its least-flooded ones are somewhat resilient to hydrological vagaries. Their production does not stop when high waters fail to reach them, meaning that they can absorb some of the water-level’s swings. This, however, only applies as long as the crisis is temporary and does not entail irreversible changes in the ecosystem.

158The consequences of the second hypothesis (a lasting change in hydrology entailing a permanent drop in water levels) are unknown. A simulation was run on the basis of a 1-metre-lower swell (high-water levels at 5.60 m instead of 6.60 m, the previous balance). This, as an aside, could occur as a result of climate change or of large facilities upstream.

159This simulation showed the following:

  • the dry grasslands growing on the delta’s outskirts and unwatered mounds could cover three times as much surface (more than 900,000 ha);

  • deep-water grass (in 150 cm to 400 cm deep water) would shrink considerably. Formations growing under 150 cm and 280 cm of water would shrivel to one-tenth of the surface they cover today.

160The deeper formations (the ones growing under 280 cm to 400 cm of water) could disappear altogether and be replaced by varieties growing in shallower water. This would hurt fodder production (these formations respectively supply 20,000 kg and 15,0000 kg of dry matter per hectare, i.e. considerably more than the delta-wide 6,000 kg average). But it would also cripple biodiversity, obliterating the Acacia kirkii forests where many palearctic migratory bird species breed.

161The drop in fodder production would be considerable: swells at 6.60 m support 15 million t, swells at 5.60 m would cut production back to 6 or 7 million tonnes a year. This area’s grazing capacity would drop 55 to 60%: the delta can support 1.5 million head a year today, but could only support 600,000 to 700,000 if this scenario materialises. Quality would drop as well as quantity: the bourgoutières, the delta’s best pastures, would shrivel. Livestock farming as we know it today, under these conditions, would not plausibly withstand this change.

The case of the water hyacinth

162Water hyacinths appeared in Mali in 1990 around Bamako and could grow all the way up the Niger River from the inland delta. In reality, however, they only thrive in certain clear-cut areas, in irrigation canals and in certain river branches near cities (i.e. where the water is “polluted” or at least contains very high amounts of nutrients). Water hyacinths are something of a hazard and can indeed cause serious concern in specific spots (dam intakes, mainly). There are biological and mechanical methods to control these plants. The Office du Niger is currently using the latter.

Fish resources and population

163From a biogeographical point of view, the Niger River’s fish fauna belongs to the Nile-Sudan province (Lévêque and Paugy, 1999). This vast province encompasses three large basins (Nile, Chad and Niger) as well as a number of smaller ones (Senegal, Volta and other). It has kept a remarkably homogeneous fauna, due to exchanges with other basins. The sheer size of this biogeographical province explains the wealth of fish varieties it harbours – and, in particular, the abundance of varieties in the Niger River. Daget (1954) argues that there are between 130 and 140 species in the Upper Niger and its inland delta.

164However, half of those species are relatively unknown (or at least small enough to slip through fishing equipment most of the time). These species are only caught incidentally or accidentally, not deliberately. So, as they can not really count as the fish resources, describing their current condition or any associated trends would be pointless. Only researchers specialising in taxonomy, studying fish stocks by poisoning small areas of the river or its branches, can supply information about the presence of these species – and can obviously only do so from a very local perspective.

165The information below, therefore, only includes the species that fishers catch (or caught). There are about 60 of them. In other words, if the question is whether biodiversity is dwindling, we cannot provide a conclusive, unbiased answer. If the question is whether some of the species caught 30 or even 60 years ago have grown rarer or disappeared, however, we can.

166Available information shows that the specific structure of the Niger River’s fish resources has indeed changed slightly since the 1950s and even since the late 1960s, in the delta and across Mali (Quensière et al., 1994; Wetlands, 1999). Very few species have disappeared (and the ones that have, no doubt, have only disappeared in specific areas). But a number of species have grown considerably rarer, and a number of other ones have gained in predominance. More specifically, we can see that one or two of the species that have disappeared or practically disappeared in the delta region (Epsetus odoe and possibly Arius gigas) did so a relatively long time ago, between the early 20th century and the 1960s. However, four species or groups of species (Gymnarchus niloticus, Parachanna obscura, Heterotis niloticus and three Polypterus varieties) have seen their ranks dwindle sharply since the 1960s, apparently in step with disappearing environments (the tall wild grass and marsh belts skirting permanent lakes). Fishers, conversely, have been seeing more Clarias and Chrysichthys in their catches. In volume terms, cichlids (tilapia, broadly speaking) predominate in fishers’ nets (alongside Clarias and Labeo). Lates niloticus are also found, if less abundantly.

167The demographic structure of these populations has changed as well. They are all younger: three in every four of the fish in an average catch are “0 +” (less than one year old) and relatively small (Laë, 1995). Bigger fish (over 3 years old and more than 50 cm long) have become very rare in most parts of the river, especially in the inland delta. However, there are still large fish in areas where fishing pressure is less intense (in the upper basin’s water courses, for instance).

168The loss of certain types of natural environments explains the changes in population makeup. But demographic changes arguably impact fishing pressure as well.

169However, these different structural changes are not, at least for the time being, affecting this resource’s yield (i.e. its ability to produce as much biomass as it loses in a year). Fish production (in tonnes a year) dropped considerably during the long drought (1973-1993) but rose again once the average or good swells returned. Production invariably moves in step with the size of a swell and ensuing flooding (see fig. 8 and 9, “Fishing” p. 200-204). Statistics allow scientists to forecast a fishing campaign’s yield based on the size of the flooded area or on the river’s flow at a given point upstream. Laë (1992b) found a link between catch sizes and an index based on the two previous floods (y1, here, is the current year, y0 the previous year and y2 the coming year):

170Annual catch from Julyy1 to Juney2 = 780.95 (flooding index)y1 + 770.71 (flooding index)y0 + 32,304

171The periodically-announced “resource drops” are indeed only transitory phenomena due to a particularly bad swell (and generally belied the following year). They may also be associated with the assessment methods: scientists, for example, may focus most of their attention on catch per unit of efforts (a ratio-type indicator which naturally drops when the effort increases).

172The developments we have discussed, however, do have negative consequences:

  • from an ethical standpoint, the fact that original bio-ecological varieties are dwindling or disappearing is regrettable;

  • there are also economic losses linked to excess fishing capacity (the amount of fish caught in a given year is inexorably pegged to the swell and ensuing flood whereas fishing capacity is inflexible) and to the drop in average fish size (i.e. their specific value per kg).

173Lastly, it is important to remember that healthy fish yields hinge on healthy river and plain ecosystems and, in particular, on generous annual flooding in the plains.

Wildlife populations and resources

174This component is especially interesting in the Niger River’s inland delta, a large area which encompasses an original and rich array of natural water and marshland environments. Wildlife elsewhere in Niger River’s valley in Mali is less exciting: it is akin to that in the neighbouring Sudan-Sahel (barring the fish, of course). This wildlife, however, is relatively poor. Pervasive farming, clearing and hunting and the last long drought have taken a toll.

175The Niger River’s inland delta harbours a wide variety of wetland habitats (marshes, lakes, ponds, flooding plains and flooding forests). Fish, reptile, mammal and water-bird colonies find food and suitable breeding conditions in these habitats. Some colonies live here throughout the year, others only stay for a season. There are other wetlands in the Sudan-Sahel area, including the Senegal delta (Djoudj), the Hadjia-Nguru floodplain in Nigeria, the Logon River’s, Chari River’s and Lake Chad’s floodplains, and the Sudd swamp in Sudan. But the Niger River’s inland delta is West Africa’s largest. Only the Okavango marsh in Botswana (in Southern Africa) and Sudd, compare.

176However, this delta’s biodiversity is far from intact. The last thirty years have seen dwindling flow rates compounded with human activities put growing pressure on the natural environment and hitting large wildlife species.

177In particular, the large carnivores that roamed parts of the upper delta and shores have disappeared. Panthers and lions are two examples (even though there may still be a few lions in neighbouring Gurma). Large reptiles (Nile crocodiles, pythons and monitors) have become very rare as well, as have large herbivorous mammals (manatees and, to a lesser extent, hippopotami).

178The Niger River’s inland delta, however, is still one of West Africa’s vital water-bird havens. Water birds nest there, rest there during their migration, or settle there for the rainy season. These birds are of afrotropical or paleoarctic varieties. The comparative size of water-bird colonies varies from one season to the next and from one year to the next, depending on the size of the swell.

179Different programmes have shown different pictures of developments in the delta. But some trends are nonetheless clear. And possibly worrying. No fewer than 125 water-bird species would alight in the delta in the early 1980s. The Wetlands project, however, only counted 103 in the late 1990s. The large varieties (giant herons, jabiru, white storks and black-crowned cranes), in particular, seem to be dwindling in numbers or presence. This trend is not as clear-cut in the case of smaller species. The Plegadis falcinellus population, for example, may well have halved over the last 20 years, whereas other species have seen their ranks remain stable.

Table 16 – Reproducing population (the number of nesting pairs), nesting water birds in the delta’s flooding

Table 16 – Reproducing population (the number of nesting pairs), nesting water birds in the delta’s flooding

Note * Annual average based on observation over a two-year period Source: Wimenga et al., 2002
Source: Wimenga et al., 2002

180The presence of large water-bird colonies seems linked to the integrity of flooded Acacia kirkii forests. Forest-clearing and poor swelling in the 1980’s and 1990s, however, were a threat to those forests. Healthier flooding after 1994, paired with efficient protection measures (Dentaka and Akkagoun forests) may have reversed this trend.

181Birds play a sizeable ecological and economic role. They protect plantations by eating locusts, grasshoppers and other insects, and thereby balance the delta’s ecosystem and bolster its production. There are no doubts about the connection between rice production, locust attacks and bird-population size.

182There are more birds during good years (the ecological niche is bigger and food more abundant). When this happens, these birds provide an added source of proteins for the population: thousands of them (teals, especially) are caught and then sold in different markets every year. Lastly, water birds are part of the delta’s appeal – and what make this and the Dogon area Mali’s leading tourist destination.

183Like Africa’s other large wetlands, the Niger River’s inland delta has been included in an international Ramsar convention. This 1975 agreement defines and organises the natural area preservation terms.

Conclusions on the Niger River’s delta and its ecosystems

184The Niger River and its valley have not harboured virgin ecosystems – or indeed any ecosystem sheltered from humankind’s hand and its effects – for a long time. A number of weather-related and anthropogenic factors over the last century have taken a toll on biodiversity in these riverside ecosystems. Neither the Niger River in general nor the delta in particular can any longer be considered as large wild, unspoilt areas. Scattered pollution phenomena have furthermore appeared, and may become threats in the future.

185However, despite these developments, the basic ecological mechanisms underlying this river’s productivity and its ecosystems (in terms of their ability to generate adequate biomass quality unaided) are intact. They are also as remarkable (from a quantitative standpoint) as ever, compared with other natural environments elsewhere in Africa and around the world. These mechanisms start working effortlessly as soon as enough water feeds them.


186Community rules and customs have shaped the way in which societies living in the Niger River valley have administrated this area and its resources for a long time. A grasslands (leydii) code established in the 19th century by the Peul Empire of Massina formalised the rules by which Peul clans administrated land. Unwritten rules enforced by “water masters” have dictated the way in which fish resources have been used for a long time as well. These traditional rules have been studied and explained by several authors after J. Daget and A. H. Ba (1955), and Gallais (1967). Political, economic and environmental developments in the second half of the 20th century have considerably altered and destabilised these systems. But they are still solidly rooted in people’s minds and behaviour.

187This chapter reviews the “modern” institutions and codes. The first such codes date back to the early days of independence (and replaced traditional systems). The second ones stem from the more recent decentralisation drive and its attempt to reconcile modern management rules and traditional local authorities.

188The institutional reform that has been unfurling in Mali since 1991 has now grown to encompass a new approach to water-resource management. This new approach can only be understood in the international realm, as the Niger River basin spans several countries.

The international framework

IWRM or a “shared vision”

189A series of international initiatives on water management and governance – born in the 1998 Paris Conference and reasserted in the 2002 Johannesburg Earth Summit – have been steering water policy at both national and regional levels.

190This new water policy, IWRM (Integrated water resources management), defines itself as a process to promote coordinated water, land and associated-resource development and management. This policy aims to maximise economic and social wellbeing and to share its benefits fairly without compromising the future of the vital ecosystems.

191IWRM also defines itself as a “shared vision” and maintains that integrated tributary-basin management is a sine qua non for proper water governance. International organisations, funding agencies and member states have agreed to focus on large trans-boundary river basins. The Mekong River basin is one such example in Asia. Examples in Africa include the Lake Chad and Lake Victoria basins, and the Orange, Volta and Niger river basins. Mali’s authorities have enshrined IWRM as official policy. Implementing it, however, will entail deep-reaching legal and institutional reform. Doing so should involve moving water management from technical and sector-specific segmentation into an integrated programme covering each distinct basin and sub-basin.

192Efforts to implement IWRM in Mali can also lean on recent (and radical) developments in legislation pertaining to water management, in the agencies in charge of administrating water, in the relationships between these agencies, and in their relationships with society as a whole.

An international institution: the Niger Basin Authority (NBA)

193This is the main international organisation. It counts nine countries skirting the Niger River (Cameroon, Côte d’Ivoire, Benin, Guinea, Burkina Faso, Mali, Niger, Nigeria and Chad) and was founded by an agreement signed on 21 November 1980 in Faranah, Guinea. It is headquartered in Niamey and aims to promote member States’ cooperation and to integrate power-generating, hydraulic, farming, livestock-farming, fishing, fish-farming, forestry, logging, transport, communication and industrial operations across this basin.

194The NBA’s specific goals follow:

  • harmonise and coordinate national policies for the development of the water resources in the basin;

  • promote and participate in development plans, by preparing and implementing an integrated basin development plan;

  • promote and participate in efforts to create and operate shared facilities and projects;

  • control and regulate all forms of navigation on the river, its tributaries and its sub-tributaries;

  • take part in applying for funding and in directing funding to studies and other work required to develop this basin’s resources.

195The NBA stays abreast of country-specific development plans. Member countries are required to inform the NBA’s Executive Secretariat of all plans touching on the basins (and in particular of schemes that might alter water availability). Member countries have also agreed not to build, on their terri-tory, schemes that might pollute water or damage fauna and flora biology anywhere along the river, its tributaries or its sub-tributaries.

The legal and institutional framework in Mali

Codes in Mali

The Code domanial et foncier (Domains and Land code)

196Water, in this code, falls into one of two domains:

  • natural public domains, mainly encompassing navigable or floatable watercourses and their banks (25-metre thoroughfares), non-navigable and non-floatable sources and watercourses, lakes, ponds, etc.;

  • manmade infrastructure, encompassing inland ports, navigation canals, irrigation and drainage canals, aqueducts and associated buildings directed to the public benefit.

The Code de l’eau (Water Code) codified by Law n° 2-006 passed 31 January 2002

197This law lays down the rules for using, preserving, protecting and administrating water resources. It applies to all the water under the hydraulic domain, i.e. to the State’s public hydraulic domain and the territorial-community public hydraulic domain under the Code Domanial et Foncier. The State is in charge of overall water management. It is responsible for administrating this resource in a sustainable and equitable manner, and for protecting it. The State may, however, transfer some hydraulic resources to decentralised communities. The ministry in charge of water has drawn up a master plan to govern water management and schemes for a 20-year period. This master plan sets qualitative and quantitative usage, development and protection targets for water resources and aquatic ecosystems. This code, in turn, lays the foundations for new water-management regulation and endorses the agencies in charge of administrating water resources. It recommends setting up a development trust within the public-sector’s water-management organisation and established a national council, regional councils, local councils and basin committees to share view and proposals regarding water-resource management and development projects. These councils and committees share their views on resource management, schemes, conflicts if and as they arise, and other issues, at their specific level.

The Charte pastorale (Grasslands Charter) codified in Law n°004 passed 27 February 2001

198The Charte pastorale lays down the basic principles governing grasslands and herding. It lays down basic herder rights (herd mobility and access to resources) and responsibilities (protecting the environment and other people’s property, in essence). Grasslands (pastures and bushes), community bourgoutières, salty land, watering spots and shelters come under State or territorial-community domains. Users are free to access these areas and facilities, but rules pertaining to how and when herds may do so apply (and are based on local customs and rights). Territorial communities administrate grassland resources in their area. They are in charge of herd trails and may regulate access to resources and charge herders for using them.

The 1995 Code des collectivités territoriales (Territorial Community Code) and related laws

199This Code des collectivités territoriales acknowledged that communities have the legal status, financial autonomy and authority to settle environmental-protection, domain-related and land-related issues (after deliberation). Territorial communities can, in particular, build and run rural or city water-management facilities, organise rural endeavours and agro-sylvo-pastoral production. If public and private domains overlap in a given community, the State may, in the public interest, transfer management rights over part of said natural or manmade domain to that decentralised community, entrusting it with its preservation.

Law n°96-050 on the principles for instituting and managing territorial-community domains

200A territorial community’s natural public domain comprises the appurtenances of the State’s public domain which are on the community’s land (and provided the State has transferred conservation and management responsibility as appropriate). These domains can include water courses, marshes, lakes, ponds, protected areas, and natural sites that have been included in the public domain by law.

201The manmade public domain encompasses buildings and other constructions erected in the interest of regions, circles or communes.

202Territorial communities are responsible for managing and furnishing their domain, and for protecting its ecological balance. They prepare land-management plans mapping out farming, grassland, fish-farming, wildlife, mining, housing and other areas. They review and – if and as required in the light of national plans – amend their schemes periodically.

203Community farming domains encompass dry-crop or rain-fed areas, irrigated areas and the associated hydraulic schemes, receding-flood areas, market-gardening areas, land that has been fallow for less than ten years, etc. Territorial communities set the levy for irrigated areas after consulting the Chambre régionale d’agriculture (regional chamber of agriculture).

204The grassland domain encompasses pastures, land that has been fallow for over ten years, herd trails and water points. Appropriate community organisations work with professional organisations and technical agencies to regulate these endeavours, and set the levies for accessing pastures after consulting the Chambre régionale d’agriculture.

205The fish-farming domain encompasses hydraulic and fish-farming facilities that communities build on their territories, and the public water areas entrusted to those communities by the State. Here, again, authorities consult professional organisations and charge levies for fishing permits.

The political institutions, administrative authorities and associations

206Mali’s Ministère des Mines et de l’Eau (Ministry of Mines and Water) is the main authority governing the Niger River in Mali. It has statutory and inherent rights over the entire water sector. It works through the DNH (Direction nationale de l’hydraulique/national hydraulics authority). The DNH is in charge of taking stock of and assessing hydraulic resources across Mali. It reviews every project that involves the water sector and supervises it directly. Ghenis (for Gestion hydro-écologique du Niger supérieur/upper Niger hydro-écological management) is one example.

207DRHEs (Directions régionales et sous-régionales de l’hydraulique et de l’eau/regional and sub-regional hydraulics and water authorities) are DNHrun decentralised agencies. They support and lend advice to territorial communities, serving concurrently in supervisory and regulatory capacities.

208The Ministère des Mines et de l’Eau also has advisory committees. These committees are responsible for consultation-based basin and sub-basin management. In that capacity, they can voice proposals on the management of the hydrographical resources and aquifer systems. They can also suggest that authorities review water-management and facilities in hydrographical basins, sub-basins or aquifer systems.

209The Comité du bassin du Niger supérieur (Upper Niger Basin committee), for example, covers the regions of Sikasso, Koulikoro, Segu and Mopti. Users, government representatives and territorial-community representatives sit on this committee, which is responsible for giving its opinion on how the basin’s water resources are used. Officially, this committee has ample consulting powers. But, in practice, wields little authority as yet.

210Conversely, the Commission de gestion des eaux de Sélingué (Selingue water management commission), the agency in charge of the Selingue dam, is functional. This commission is chaired by a DNH representative. Projects can not move ahead without consulting it, and its members convene once a month. Its responsibilities encompass the following:

  • review the various water-management programmes drawn up by national organisations using the dam’s water, to ensure they are suitable;

  • decide on dam water-management programmes (filling and discharging);

  • brief decision-makers on how water-management decisions will impact the associated technical aspects (power generation, migration, fishing, health, industry, tourism and the environment);

  • keep the dam in good working condition.

Note *DNAER : Direction nationale de l’aménagement et de l’équipement rural
Note **DNAMR : Direction nationale de l’appui au monde rural
Note ***AGETIER: Agence d’éxécution des travaux d’infrastructures et d’équipements ruraux, the agency in charge of rural infrastructure and equipment
Note **** AGETIP: Agence pour l’éxécution de travaux d’intérêt public contre le sous-emploi, the agency in charge of public-interest projects and of eradicating underemployment
Figure 10. The organisations involved in water management

211The Ministère de l’Environnement (Ministry of the Environment) works mainly through the ABFN (Agence du bassin du fleuve Niger/Niger River basin agency) founded by a 2002 ruling. This agency is in charge of protecting the Niger River, its tributaries and their tributary basins. Doing so, in particular, involves the following:

  • bolstering river, tributary and tributary-basin management resources;

  • improving and streamlining water-resource management for the various users;

  • helping to prevent natural hazards, counteracting pollution and nuisances, and keeping the river navigable;

  • liaising and cooperating with counterpart technical organisations in other countries skirting the Niger river and basins;

  • devising and implementing a levy system for organisations diverting or polluting water, and administrating that system’s proceeds.

212Other ministries are involved in water management as well:

  • the Ministère de l’Agriculture, de l’Élevage et de la Pêche (Ministry of Agriculture, Livestock and Fishing) is in charge of hydro-agricultural schemes and fishing, and takes an interest in water quality;

  • the Ministère de l’Administration territoriale et des Collectivités locales (the Ministry of Local and Territorial Government) is in charge, inter-alia, of emergency measures, and of liaising with ministries in charge of foreign relations and civil defence in the event of natural catastrophes and floods. Three types of territorial organisations can have authority over water-management policy. These follow:

  • communal councils are in charge of water- and sanitation-related policies on a local level. They have councils dealing with different areas (fishing, livestock farming, etc.);

  • circle councils work with technical authorities to review community projects for approval;

  • regional assemblies make sure projects tally with national-level programmes.

213Territorial communities take part in the meetings of basin committees and councils founded by the new Code de l’Eau (Water Code).

214Lastly, associations have grown into prominent roles since the 19911992 regime change. Society as a whole, in other words, is playing an active part through decentralised elected organisations (territorial communities and their councils) and through organisations representing the various trades using the Niger River (fishing, farming and livestock farming, mainly). These associations are very involved, and consulted on a regular basis. Examples include Apram (Association des pêcheurs résidents au Mali/Mali resident fisher association), APPM (Association des pêcheurs et pisciculteurs du Mali/Mali fisher and fish-farmer association) and the Association pour le développement intégré à base communautaire (an association for integrated community-based development).

Examples of the operations-based approach

215Mali’s Constitution has enshrined consultation as a sine qua non for management decisions or policy. Consultation unfurls in three spheres:

  • the Assemblée nationale (National assembly), through its Commission du développement rural et de l’environnement (rural development and environment commission), which instigates environmental legislation alongside the Haut Conseil des collectivités (community high council);

  • the Haut conseil des collectivités is responsible for reviewing and sharing its views on all local and regional development policies (the government is required to ask for this council’s views on these issues);

  • the Conseil économique, social et culturel (économic, social and cultural council) has authority over all aspects of economic, social and cultural development. This encompasses issues such as protecting the environment and promoting quality of life.

216As far as water management is concerned – and more practically speaking – consultation will be possible through the basin commissions, committees and water-management bodies that have been established or are already operational at different levels in the Government (cf. “The legal and institutional framework in Mali” p. 227).

217The Commission nationale de gestion des eaux (national water-management Commission) is a good example. Since its inception, this agency has been working to strengthen ties between the thereunto remote organisations running Selingue and Markala. These efforts, in particular, have allowed Markala (and hence the Office du Niger) to tap regulation at Selingue more efficiently. This organisation has also set the low-water floor downstream from Markala at 40 m3/s, in the interest of users in that section of the river.

218Planning is in the hands of the various relevant ministries and of the national agencies reporting to them. Planning, here, involves defining sec-tor-specific policies in the form of master plans. All sector-specific policies are now coherent from a theoretical standpoint and in line with CSLP (Cadre stratégique de lutte contre la pauvreté/poverty reduction strategy), the country’s political beacon directing all sector-specific policy. The SDDR (Schéma directeur du secteur développement rural/master plan for rural development), has been and will be shaping rural endeavours from 2000 to 2010. Its strategy stems from an analysis of developments in the area and from across-the-board consensus (encompassing the State, territorial communities, users, associations and partner development organisations. That strategy maps out areas of responsibility and lays the foundations for development programmes in the area. Developing irrigation is a top priority. Efforts in this direction have spawned the SNDI (Stratégie nationale de développement de l’irrigation/national irrigation-development strategy). This scheme focuses principally on developing total-water-control operations and pursues six main goals, namely:

  • streamlining irrigation-infrastructure design and cutting costs;

  • involving NGOs, helping them to access financing;

  • improving irrigated-area management;

  • increasing irrigated-area production and yield;

  • reforming the sub-sector’s institutional and legal environment;

  • minimising the negative environmental and social impact of irrigation.

219Generally speaking, the SNDI’s choices aim to increase irrigated areas and to enhance production-system yields.

Monitoring and circulating information

220State agencies in charge of each environmental sectors or subsector gather, monitor, circulate and compile information – mainly by collecting data and publishing statistic reports on a regular basis – separately. Combined, they could provide a fairly good picture of the various aspects of Mali’s environment and rural sector – and in particular of those around the Niger River. The DNH gathers the bulk of the hydraulic information and stores it in Hydrom, a database that is currently being upgraded (Hydraccess). Table 17 (p. 234 to p. 239) shows the systems used to gather and monitor information about the Niger River.

Table 17 – Systems and organisations gathering, monitoring and circulating information

Table 17 – Systems and organisations gathering, monitoring and circulating information

Table 17 – Systems and organisations gathering, monitoring and circulating information

Table 17 – Systems and organisations gathering, monitoring and circulating information

Table 17 – Systems and organisations gathering, monitoring and circulating information

Table 17 – Systems and organisations gathering, monitoring and circulating information


1 Maximum high-water flow rates in Ké-Massina (downstream from Markala, at the entrance of the delta) are in a 3,000 to 5,000 m3/s bracket, meaning a 500 or 550 m3/s change would be tantamount to a substantial 10% to 18% of the flow.

2 This system is also in use in Kayes, in small areas feeding off the Senegal River.

3 Biological Oxygen Demand, an indicator of biological pollution.

4 Chemical Oxygen Demand, an indicator of chemical pollution.

Table des illustrations

Titre Table 2 – Food consumption in Mali (in tonnes)
Légende Source: Based on 1988/97 DNSI consumption budget survey
Fichier image/jpeg, 112k
Légende Source: NOS, 2003Figure 3. Increase in Mali’s power generation
Fichier image/jpeg, 124k
Légende Figure 4. The Niger River’s maximum flow rates in KoulikoroThe flow peaks in Koulikoro illustrate the differentvariability scales in the Sudan-Sahel (year-on-yearand decade-on-decade patterns, mainly), even if adam was built in Selingue in 1982.
Fichier image/jpeg, 112k
Légende Figure 5. Observed and simulated rainwater anomalies (variation %)Source: McCarthy et al., 2001
Fichier image/jpeg, 292k
Titre Table 3 – Diversions by Dams on the Niger River’s upper and middle tributary basin
Fichier image/jpeg, 60k
Titre Table 4 – The impact of these two large dams on downstream flow rates during high-water and low-water stages
Fichier image/jpeg, 148k
Légende Source: FAOSTATFigure 6. How Mali’s cereal production and population have evolved since 1961In spite of sharp year-on-year fluctuation, production has, by and large, grown faster than requirement.
Fichier image/jpeg, 300k
Titre Table 5 – Cereal production, farmland area and yields in Mali (1961-2003)
Légende Source: FAOSTAT
Fichier image/jpeg, 152k
Titre Table 6 – How Mali’s rice production has evolved since 1961
Fichier image/jpeg, 60k
Titre Table 7 – Usable farmland, traditional rice plantations, 1972, 1978 and 1994
Fichier image/jpeg, 40k
Titre Table 8 – The Niger River’s supply at Koulikoro and water diversion in gravity-governed total-water-management systems
Légende Source: Kuper et al., 2002
Fichier image/jpeg, 284k
Légende Figure 7. Size of Office du Niger areas by scheme type, crop and season
Fichier image/jpeg, 204k
Titre Table 9 – Different crops feeding off the river, their requirements, and their impact on the flow
Fichier image/jpeg, 224k
Légende Figure 8. Fresh fish landing (in tonnes)Recent changes in Mali’s fish production as reported by national agencies to FAOSTAT. The bulk of production (over 95%) comes from the Niger River basin.Source: FAOSTAT (updated in early 2004)
Fichier image/jpeg, 108k
Légende Source: Morand and Kodio, 2004, based on Observatoire de la pêche figuresFigure 9. Median catch per unit of effort (per fishing trip, in this case) against a swell index (how many days the water level stays above the 4.50-m mark in Mopti)The round dots and top line show figures for the first half of the fishing campaign (December to mid-March); the squares and the bottom line show figures for the rest of the campaign (mid-March to June). The numbers denote years.
Fichier image/jpeg, 120k
Titre Table 10 – Water requirements for fish and fishing
Fichier image/jpeg, 92k
Titre Table11 – Water requirements for cattle herding
Fichier image/jpeg, 84k
Titre Table 12 – Delta fodder production and theoretical change based on different high-water levels in Mopti
Légende Source: Marie, 2002
Fichier image/jpeg, 72k
Titre Table 13 – Water requirements and impacts, household consumption
Fichier image/jpeg, 100k
Titre Table 14 – Water requirements and needs, river sailing
Fichier image/jpeg, 116k
Titre Table 15 – Wood potential across five circles in the Niger River’s inland delta
Légende Note * 1m3 = 2.3 stere et 1 stere = 330 kg Source: SED project, 1999Source: SED project, 1999
Fichier image/jpeg, 92k
Titre Table 16 – Reproducing population (the number of nesting pairs), nesting water birds in the delta’s flooding
Légende Note * Annual average based on observation over a two-year period Source: Wimenga et al., 2002Source: Wimenga et al., 2002
Fichier image/jpeg, 228k
Légende Note * DNAER : Direction nationale de l’aménagement et de l’équipement ruralNote ** DNAMR : Direction nationale de l’appui au monde ruralNote *** AGETIER: Agence d’éxécution des travaux d’infrastructures et d’équipements ruraux, the agency in charge of rural infrastructure and equipmentNote **** AGETIP: Agence pour l’éxécution de travaux d’intérêt public contre le sous-emploi, the agency in charge of public-interest projects and of eradicating underemploymentFigure 10. The organisations involved in water management
Fichier image/jpeg, 128k
Titre Table 17 – Systems and organisations gathering, monitoring and circulating information
Fichier image/jpeg, 292k
Titre Table 17 –
Fichier image/jpeg, 264k
Titre Table 17 – Systems and organisations gathering, monitoring and circulating information
Fichier image/jpeg, 300k
Titre Table 17 –
Fichier image/jpeg, 220k
Titre Table 17 – Systems and organisations gathering, monitoring and circulating information
Fichier image/jpeg, 300k
Titre Table –
Fichier image/jpeg, 372k

© IRD Éditions, 2007

Conditions d’utilisation :