Version classiqueVersion mobile
OpenEdition Books

Avenir du fleuve Niger

Jérome Marie
Pierre Morand
Hamady N’Djim

Synopsis and Recommendations

The problems: how they are addressed today

Texte intégral

1In this section, we will focus on problems – i.e. the particularly complex issues that authorities have to tackle. In most cases, these problems arise where forcing-factor dynamics (the weather, for example), living-resource dynamics, user dynamics, social dynamics and institutional dynamics cross paths. These are the factors we reviewed separately in the previous chapter.

2We suggest classifying theses problems in four categories. First of all, we will look at hazards. Hazards are sometimes sudden, do not last long, and are triggered by random factors. People have little or no control over hazards. Second, there are tensions. Tensions include ongoing friction and overt or concealed affliction due, to a large extent, to human activities or institutions. Third, there are threats. Threats are not yet news, but they will jeopardise the future – especially if the organisations in charge of them do nothing to thwart them. Environmental dynamics – by themselves or compounded in a negative or unforeseen manner with human behaviour, play an important role in these first three categories.

3Lastly, there are what we will call “institutional difficulties” in reference to the thorny issue of improving governance systems. We expect institutions to find a way of sheltering people from environmental threats and nurturing consensus on how to share wealth.

4But, before reviewing the problems associated with the categories listed above, it is important to take a quick look at another problem: health (and river diseases in particular). It is a known fact that a number of human-pathogen viruses, bacteria and other agents (worms and protozoa), and a number of animals (insects and molluscs) that carry those agents, live or indeed thrive in tropical water environments. The people in the Niger River valley – not unlike people in the other damp Sudan-Sahel areas – thus face heavy morbidity and mortality rates. Malaria, schistosomiasis, cholera (with endemic seasonal outbreaks in the region of Mopti), leishmaniasis and carbuncle are the main causes (even though carbuncle affects cattle, mainly). Other infectious diseases, more linked to poverty and behaviour patterns than to the natural environment, take heavy tolls as well. These include trachoma, meningitis, amebiasis and hepatitis varieties, tuberculosis, and AIDS. These diseases, along with the fact that healthcare facilities are scarce in rural areas and that transporting the sick is difficult, explain the notoriously sombre health statistics. The region of Mopti, on the Niger River’s inland delta, stands out with an infant mortality rate of 325/1000 between 0 and 4 years of age against a 237/1000 national average (Coulibaly et al., 1996). A number of studies have shown that irrigation schemes can help river diseases spread in tropical regions (Samé-Ekobo et al., 2001).


Bad-weather years

The phenomenon and circumstances

5This is the biggest risk for rural populations. A lack of rainfall compounded with a poor swell is a real and practically unpredictable possibility every year. Bad years hit farming and fishing campaigns (as well as livestock farming). They are also frequent: since the 1970s, observers agree, rainfall and swells have been “fairly bad” or “very bad” at least one year in three. This weather hazard often blankets the Niger River basin across Mali. Rainfall is also poor in neighbouring Sudan and Sahel those years. Poor flow figures in the Niger River’s upstream and middle stretches correlate to poor rainfall upstream (logically) but also hinge on rainfall in the Niger River’s stretch through the Sahel (which could be otherwise, as rain in the Sahel does not “make or break” the Niger River’s swell). The result is that there are not many chances of compensating generally unfavourable conditions (poor swells and problems with flood-fed crops) with isolated favourable conditions (rain and rain-fed crops) which could, in specific places, be more favourable.

6The reverse does not apply: a given year may bring generous rainfall and a healthy swell, but the conditions for certain crops may be bad in certain places. Here, finer weather hazards are at work: how well or otherwise rainfall and swells synchronise, for instance, will make or break production in traditional rice-growing schemes.

The impact

7Bad-weather years (such as 2004) have a huge social and economic impact. The rural populations can face food shortages during the following year’s gap phase: cereal can indeed run out in May or June, when the following harvest is not due until October or November. The country’s economy as a whole suffers as well: the balance of trade weakens (Mali has to import rice and frozen sea fish) and growth sputters (or the economy goes into recession).

The management approaches

8Bad-weather years are one of the harshest chronic woes that authorities in Sudan-Sahel countries such as Mali have to cope with. They have a number of options:

  • sending food aid to help affected rural populations (and targeting it not to shake the agro-product market’s balance). This is what the World Food Programme does at the SAP’s (Système d’alerte précoce/Advance Warning System) request;
  • building large irrigation schemes to shelter parts of the farming population from these hazards (provided, of course, that these schemes can stock up enough water upstream);
  • trying to contain the impact by circulating information that will help people avert hazards a few weeks before they occur, and handle them more efficiently once they do (for example altering their annual crop strategy or investments in fishing; the Agrhymet regional system).

Exceptional swelling and flooding

The phenomenon and circumstances

9As we have said, one of the Niger River’s hallmarks is its ample swell. Swells usually build up gradually over several weeks and crest as soon as the rainy season ends. Crests then move downstream between Bamako (in early or mid-September) and Gao (in late December or January). They abate (or flatten out) as they move downstream.

10Flooding, it follows, is a routine phenomenon that comes around every year when the swell reaches each area in its turn. It includes all the banks and adjacent plains. In the delta region, this flood stretches several thousand square kilometres and lasts seven to ten weeks centred in September to November.

11Generally speaking, the livestock and people are familiar with and expect these flow and water-level cycles. They work around and move around the river’s cycles, and know how to harness them to improve their production.

12So, before looking into how floods can threaten human lives and endeavours, we have to distinguish what, within this annual regime, can be considered a threat. There are two such events:

  • swells may rise at a normal pace (gradually) but build into exceptionally high crests;
  • swells may rise faster than usual (over 5 cm a day) at a given point or on a given reach.

13Then we can analyse the circumstances that promote these events. Here, two or three fairly different events are worth pointing out. These follow:

  • type A events are actually the opposite of poor-weather years: they are years that bring excessive rainfall. DNH frequency analyses have linked swell levels to recurrence patterns. There are decade swells (one, on average, every ten years), century swells and so on. This gives a fairly good idea of the chances of such a hazard occurring. However, definitions such as these are incomplete without another factor: generation perception. The 1994 swell hit the 6.44 m mark in Mopti. That was perfectly normal by the century’s statistical average. But people remember it as exceptional because it followed 20 years of very poor floods. Young people had never seen swells reach those heights, and their elders no doubt had some trouble remembering the ones that had. Comparable floods in 1999 and 2003, of course, were not perceived that way;
  • type B events can be linked to particular meteorological circumstances. Very abundant rainfall in the Niger River’s upper basin at the end of the rainy season builds up a 15 to 40 cm “swell peak”. This peak can last a few hours or several days. It rides the river’s seasonal swell wave which is also at its crest. If Selingue is already full (at maximum security level), it can not absorb the water from Sankarani, so that swell merges into the Niger as well. These swells travel very fast down the river to Bamako or Koulikoro, and cause some damage on the banks. But the crests do not go very far: they dwindle as soon as they reach the low banks and flooding plains that spread them out sideways. These phenomena probably occur between 10 September and mid-October (as in 2001). Deforestation along tributaries exposes soil to stripping, accelerates flow rates, and exacerbates this phenomenon.

14Different circumstances (Type B2 events) can also cause water levels to rise very fast: incidents or accidents in hydraulic facilities (the worst-case scenario, of course, is a dam failure). In these cases, swell crests can occur any time of year. They may not be high, but they are sudden. That is why they can damage facilities or wreck human endeavours. If Markala spillway gates open too fast during the low-water stage, they can cause flooding as well. Again, these swells only rise about 10 to 20 cm, but their effects are noticeable quite a way downstream (as far as the delta). This is because, during this season, the river is in its narrow bed (which concentrates and propagates waves).

The impact

15Type A events do not carry much of a risk for rural populations and endeavours: people there have always known better than to settle in the lower parts of the valley and delta. Traditionally, they have only built lightduty dwellings using plants and trees, in what are called campements (camps). These campements usually only stand for a season, as dwellers desert them when the swelling comes round again. Some, however, can become permanent banco (dry-sludge) houses when poor flooding stretches over a number of years. They would still be fairly easy to evacuate should the need to do so arise: the people living in them (fishers and rice farmers) usually have pirogues. Other people living in rural areas are less prepared, but lead sedentary lives in villages on high mounds (toguere) which never flood.

16This is not the case with urban and – especially – suburban populations: people leaving their villages and moving to city outskirts may not have a chance to get to know the area before unknowingly or unwillingly settling in flood-prone dips or bas-fonds (the name of a neighbourhood in Mopti, and the French term for a dip and for a shanty town). Unusually high swells can hit these people hard: they could take their belongings, contaminate their wells and flood their neighbourhoods, turning the area into health hazards and allowing diseases to spread. Such floods could also hit industrial or storage facilities, damaging them and possibly releasing noxious chemicals into the river.

17In Mali’s stretch of the Niger River, type B and type B2 events (sudden natural or artificially-caused water-level surges) have not, as yet, reached catastrophic proportions. Indeed, in the upstream reaches, around Bamako, a fast (relatively speaking) rise in water levels can wash away fishing boats and tackle (about 100 were lost in September 2001), surprise riverside-warehouse caretakers (who might not be able to move the merchandise into a safe place on time) or surprise staff running takeouts in small irrigated areas (meaning the plots might flood too fast). But, at the end of the day, the effects are limited and have not, apparently, caused the loss of human life. These phenomena are not particularly frequent, either (they occur once every five to ten years). The biggest risk, again, would be that the flood poured large amounts of poorly-stored pesticides or hydrocarbons into the river. If this occurred on a large scale (if a large dam wall broke, for instance), it would obviously be much more serious and could threaten human life.

18In any case, the risk of flooding (i.e. the chances of a flood occurring and its breadth if and when it did) has not changed at all over the years. What have risen sharply are the associated risks: more people are settling right by the river. Fast population increases in cities and Mali’s economic development of course explain much of this trend.

Management approaches

19There are three aspects: prevention (involving zonation, counteracting erosion and regulating swells), fast warning capabilities, and the ability to deal with a catastrophe when one occurs.

20The best way to prevent all flood-related risks (whatever the type) is to limit human dwellings in flood-prone areas. This involves splitting an area into zones and mapping risk levels in each zone. Mapping out the chances of flooding in each zone, however, would not be enough. It is important to map out how floods might occur in each zone (whether they might strike suddenly or build up over a period of time, when they might occur, etc.), in particular in reference to A, B and B2 events. In some high-risk areas, factories or warehouses should not be allowed, under any circumstances, to use or store products that could pollute the river. This involves enforcing zonespecific rules once they are written down. It will be especially difficult to do so in towns, where pressure on land is high. As an aside, mapping out rural zones should not entail banning fishers, rice farmers and breeders from setting up seasonal camps in low areas (they build these camps to be near their work at that time of year).

21Counteracting erosion in the upper basin by limiting flow speeds could reduce the chances of Type B events.

22Preventing and regulating floods is difficult – or perhaps indeed impossible – in Mali at this point in time. Selingue is responsible for keeping the river’s level under 370 cm in Bamako to avert catastrophes. It should also be able to absorb sudden swells from exceptionally copious rainfall in the upper basin until the beginning of September (i.e. as long as the lake level does not get near its maximum security level, 349.00 m). But, in practice, the river’s seasonal crest reaches Bamako on 10 or 15 September. At that point, Selingue is full (or almost full), and can not absorb the swell. That is when a high-water crest could be dangerous for the city. So Selingue’s role mitigating unexpected swells is very limited. The September 2001 flood (which Selingue was unable to contain) proves this point. And this dam, it is worth adding, is on one of the Niger River’s tributaries.

23Early-warning systems are important mainly inasmuch as they take the surprise element out of sudden flooding (Type B events), by giving people and businesses time to prepare. They would also, of course, play a critical in the case of a serious event (caused by a broken dam). Using government and military radio stations as well as national and local radio stations could save lives.

24Containing a catastrophe once it hits involves limiting the human, economic and environmental consequences. Current plans involve the following:

  • providing for the populations’ needs. This includes providing emergency accommodation (in classrooms), securing water supplies (protecting wells to avert health hazards), moving people to safe areas, providing temporary living quarters there (and providing light, water supplies, etc.), providing food for at least three months, and providing medical care for victims;
  • draining or pumping stagnant water out of flooded homes, villages, towns and plots;
  • assessing the impact of the floods on production and providing for the population’s subsistence accordingly (deferring loan repayments, supplying 900 kg of grain for each hectare of forgone harvests, preventing epidemics that might affect livestock).

Water pollution

The phenomenon and circumstances

25Other circumstances – most often involving human activity – can also trigger serious pollution. Examples follow.

  • Wastewater-treatment incidents in the Huicoma plant in Koulikoro. This plant usually only treats about 40 to 60% of its wastewater and pours the rest into the river. When its system fails (as is often the case), it pours all its wastewater into the river.
  • Wastewater-treatment incidents in the Siribala plant. If the backflow pump breaks down, water pours into the Office du Niger’s arterial drainage, momentarily polluting the water there.

26There is also a risk of massive hydrocarbon spillage causing exceptional pollution (incidents while handling storage tanks or accidents involving fuel tankers are the biggest risks).

27Massive rainfall can hit intensive-farming areas (cf. “Agricultural pollution in the river” p. 199) and also cause severe pollution. Such phenomena seem to occur in the upper basin during the rainy season, causing isolated fish mortality in the Selingue impoundment. In 1993, 1997 and 1998, heavy rain caused DDT pollution in the river by Bamako (whether the DDT came from farms is not yet entirely clear). This episode only lasted a few days but was serious enough to disrupt life in that city.

28Lastly, certain hydro-geological configurations can promote pollutant migration to extremely sensitive areas. Bamako’s deep groundwater table, for instance, is fed by the river through a fault. It could, therefore, be exposed in the case of severe pollution.

The impact

29All in all, severe pollution episodes to date have been fairly small in size and mainly only hit fish. In Koulikoro, for example, waste-related incidents have killed fish several kilometres down the river. In Siribala, fishers have seen that high pollution concentration can cause instant fish asphyxiation. This deadly effect is noticeable about 20 km beyond the pollution source. How long it lasts is not yet clear but, according to fishers, it can last about one month.

Management approaches

30Handling these incidents involves prevention (through regulation and by improving treatment systems, essentially), and detection and warning capabilities (reaching the populations that consume the water). When DDT pollution was found in the river’s waters in 1993, 1997 and 1998, the population and EDM (Énergie du Mali/Mali’s power utility) network users were warned through the press.

Pest proliferation

The phenomenon and circumstances

31Insects (locusts and grasshoppers), rodents and small birds proliferate often in the Sahel region, and assail the Niger River valley and inland delta (as well as other regions). The causes triggering this phenomenon are varied and complex: they hinge on ecological processes and on climatic factors. Locust proliferation has been studied in depth, and ranks among the most spectacular such phenomena. The last locust proliferation stretched from August to October 2004. One of this phenomenon’s hallmarks is that it generally occurs after a string of good-weather years, i.e. years bringing adequate rainfall in the Sahel and in other areas skirting the Sahara (in the Maghreb). Proliferation of harvest-devastating small birds, conversely, tends to be more regular from one year to the next.

The impact

32The impact of these proliferation phenomena is akin to that of a badweather year. Pest proliferation usually occurs on a smaller scale but nonetheless takes a toll on harvests and can cause a food shortage the following year.

Management approaches

33Several organisations and several national and international programmes (Agrhymet, PICDCS, FAO) play an active role trying to forecast, forestall and thwart predator proliferation. An agreement that will involve coordinating and running a sub-regional scheme to counter grasshoppers (Cosrela) was recently drawn up.

34In the case of a confirmed catastrophe, food aid (WFP) can promptly be made available for the targeted populations in the hardest-hit areas, through the advance-warning system (SAP).


35Two of the tensions that can degenerate into open conflict or, conversely, that appear as persistent but only barely perceptible or indeed silenced friction, stem directly from sharing water (cf. “The impact of hydraulic schemes on flooding and flood-dependent production systems” p. 250), wetlands and other areas near the water (cf. “Tensions between farming and transhumant livestock farming over sharing land during the flood recession and dry season” p. 255).

The impact of hydraulic schemes on flooding and flood-dependent production systems

36The size of flooded areas (and therefore the height of the swell wave), as we have seen, has a strong impact on production systems and users (cf. “River-usage characterisation and performance, water requirements and relative impacts” p. 184). This is principally the case in traditional (unaided) rice-farming areas, rice farming in controlled-flooding areas, fishing, and itinerant livestock farming. None of these systems are very productive (and three of them can be fittingly described as “extensive”). However, they can share the same area, using it in turn as seasons go by. Typically, a hectare of flooding plains can produce 100 kg of traditional rice (because crop areas only account for 10 or 15% of the flooding plains), 50 to 100 kg of fish and 7 to 10 kg of meat a year. That hectare of flooding plains does not produce that amount of meat alone (herds spend part of the year on Sahel pastures). But it is fair to say that, without the flooding plain and the fodder it produces from December to June, meat production would almost or entirely disappear. Fish production is similar: fish are often caught on river beds, but grow up in the plains. Without those flooding plains, the river would not, strictly speaking, produce much. In some cases (fish, for example) the productivity ratios provided above can be confirmed by statistical links between changes in regional production and changes in the size of the flood covering these plains. All this goes to say that flooding plains are environments that produce generous and measurable amounts of food. These environments supply most of Mali’s fish.

37It is also important to remember that these traditional flood-dependent production systems provide livelihoods for a large population. If we assume that three-quarters of the people living between Ke Massina-Djenne and Timbuktu live off one or more of these endeavours or off processing or selling the produce, it is plausible to say that 900,000 people live off these activities.

38How upstream hydraulic schemes might impact flood levels has to be understood in this light. As we have seen, existing schemes may only have a modest impact. But they are not inconsequential. The gradient linking swell levels and flooded delta areas is well known, at least for the area south of Lake Debo (fig. 11): in common swell brackets (between 5.60 m and 6.40 m in Mopti), a 10-cm rise in water levels (again, in Mopti) floods about 980 km2. of land. A 10-cm drop in water levels, it follows, will have a sizeable impact on production systems in the plains.

Source: Marie, 2002
Figure 11. Flooded areas in the delta (up to lake Debo) paired with swell levels in Mopti

39Some authors have gone further and suggested ways of calculating production losses in the delta. One (Laë, 1992a) has attempted to peg fish-production losses to water-level drops due to upstream swell diversion. Speculative mathematics aside, the data provided here seems to be enough to contend that river management and schemes have to take the needs of production systems in flooding plains into account.

40To date, these requirements – mainly expressed in terms of swell levels and flooded areas – have not been properly factored into scheme-opportunity surveys or into scheme-management guidelines. Moreover, figures supplied by official hydrological agencies (which focus on water quantity and flow rates, never on water levels or flooded areas) only provide total annual diversion or instantaneous discharge rates during low-water stages. That information is irrelevant to swell management.

41Lastly, ample natural flood variation (from one year to the next and from one decade to the next) makes it difficult for delta populations to make out exactly how much water they lose to diversion and storage upstream. Society as a whole has not broached this question in enough detail. However, some communities (fishers, especially) have long accused the large dams (Selingue in particular) of wronging them.

42All this explains why the issue of how to distribute the Niger River’s swell water in Mali is not addressed at the right level today. But new schemes in the upper basin (Fomi in Guinea, for example) and possible Office du Niger extension plans should provide an opportunity for experts and for Mali’s society as a whole to broach this question.

User friction over low-water flows

43A look at requirements associated with different endeavours and uses (cf. “River-usage characterisation and performance, water requirements and relative impacts” p. 184) shows that low-water stages (February-June) spark fierce competition for the scant available water. This competition, broadly speaking, can be described as follows:

  • on the one hand, diversion into large gravity-governed areas during the low-water stage (Office du Niger mainly, and Selingue and Sotuba far behind) to irrigate off-season crops (market gardening and a little rice, from February to May) and to start watering rice beds for the coming plantation during the rainy season and swell;
  • on the other, the needs of users downstream. These needs span the whole spectrum: pump-irrigated village plantations, smack navigation, household consumption (for riverside population all the way down to Niamey), maintaining at least minimum water levels for fish-resource survival, and drinking water for cattle.

44The water released from Selingue upstream eases some of this tension with a 70 to 120 m3/s boost to natural flow rates during this season. This typically keeps rates at or above 40 m3/s after Markala (on Commission nationale des eaux instructions).

45This management approach works fairly well inasmuch as it keeps flow rates at more or less the same levels as stations in Mopti recorded in the same season in the 1950s and 1960s. These minimum flow rates seem to be enough for the ecosystem (for fish, especially). However, some users downstream – water consumers in the city of Niamey, especially – feel that the shortage is worsening – either because the number of consumers is increasing or because the small amounts of water diverted here and there up the delta (and pumped into small village irrigation plots) have finally started making a noticeable dent in the flow. Whatever the case, historically, the Niger River has always had severe low-water years in Niamey.

46To minimise water diversion during low-water stages, Office du Niger would like to cut back off-season rice production (or stop it altogether) and plant crops that use less water instead.

Poor scheme water efficiency

47As we have said (“Agriculture”, p. 190) large hydrologic schemes (such as the ones that Office du Niger runs) do not use water efficiently enough. Firstly, the networks waste a lot of water.

48First of all, there are huge differences between flow rates in Markala and flow rates entering the primary canals: data collected in the 1995-1996 campaign show that nearly 40% of the 36 000 m3/ha that enter the scheme are wasted. Large pipes and falas (infiltration, evaporation and discharges to regulate water levels) explain some of the loss. Off-bed plantations explain another portion (water supply, in this case, can not always be controlled). But efficiency in second-tier and third-tier networks is also a problem: some divisors and irrigators supply less water than an area theoretically needs, and others supply twice as much water as necessary.

49Based on PSI (Pôle système irrigué/irrigated system unit) research, we can summarise the situation as follows.

  • Water loss in hydraulic systems upstream of the Office du Niger (Sahel canal, Molodo fala, etc.) account for 40% of the total volume delivered by Markala. In other words, 60% of the water reaches the irrigation canals and beds.
  • Of the 60% that reaches irrigation channels:
    • 20 to 25% (i.e. 12 to 15% of the total delivered by Markala) is lost to infiltration in the irrigation canals themselves, especially off season;
    • 35% (i.e. 20% of the total delivered by Markala) is lost to bed drainage during the rice-farming season;
    • 40% (i.e. 25% of the total delivered by Markaka) is actually used by plantations.

50During the rice-farming season, water consumption using Retail dispenser valves stands at 14,700 m3/ha (in line with the Office du Niger’s 15,000 m3/ha target). Factoring in drainage loss and rainfall, each hectare receives about 12,000 m3 on average. This is enough for rice plantations (Beau, 1981).

51The results, however, mask the sharp differences between regulation systems and how easy or otherwise they are to run. Regulating divisors equipped with facings (Grüber beds) seems to be the most efficient option (in spite of the fact that they have control valves in the irrigating nozzles, meaning that they supply less volume at a more regular pace). Irrigator regulation with mask dispenser valves (Retail) is not efficient inasmuch as the instruction levels are fulfilled in the divisor. Regulating divisors using flat valves and irrigators with part modules (Kouya beds) is the least efficient option. This variability increases at the divisor and at the sprinkler level.

52Out of season, the 5,850 m3/ha of water that enter the scheme are far higher than plantation requirements, which only represent about 20% of the area using Retail divisors. Watering the network entails losing an estimated 1.5 mm per day to infiltration. This water builds up in groundwater tables and rises back to the surface by capillarity. When it does so, it brings salt with it. This damages paddy fertility.

53Another way of assessing water efficiency in a given area is looking at the way in which costs are distributed and at whether they could give users a greater sense of responsibility. At this point, the levy that farmers pay only covers primary and secondary network maintenance and water-management services: it does not cover the amount of water they actually use. One way of improving efficiency could involve new management schemes that include – in particular – pegging levies to water consumption.

54Generally speaking, improving water efficiency in these areas is a toppriority Office du Niger goal and part of SNDI’s broad agenda. This is also one of the goals of international development projects such as Appia (Amélioration des performances des périmetres irrigués en Afrique/improvement of performances on irrigation schemes in Africa).

55To a certain extent, we can hope that extending irrigated areas will in itself spur greater efficiency. Stretching irrigated areas will involve stretching water resources, and therefore entail administrating it more efficiently, even if doing so involves improving network and paddy upkeep, improving infrastructure, and improving work-organisation and planning.

56Whatever the case, implementing the various technical and management instruments to improve efficiency seems to be a sine qua non in efforts to extend irrigated systems with equanimity (which, in turn, is a political goal). The impact that these systems have on high-water and low-water flow levels has to accommodate users downstream. Diverting substantially more water into irrigation schemes will necessarily involve building a new storage and discharging dam upstream to support low-water flows and to keep them at the prescribed 40 m3/s downstream from Markala – which, at the end of the day, would be tantamount to diverting more swell water.

57The Office du Niger’s new plan to increase water efficiency involves cutting back (or even banning) off-season rice plantations and using areas now operated out-of-bed for extensions first.

Tensions between farming and transhumant livestock farming over sharing land during the flood recession and dry season

58In 1989, flood-fed rice plantations spanned 160,000 ha in the delta (“Agriculture” p. 190). These plantations were nomadic: farmers moved their paddies according to where they believed the swell would benefit them. The last 50 years have seen these crop areas grow in tandem with the population. Poor flooding in the 1970s and 1980s drove rice farmers to deeper basins, where they cleared the inland delta’s best pastures (the bourgoutières, in particular).

59Back in 1952, bourgoutières were only rarely cleared. By 1989, however, 51,252 ha had been cleared and turned into rice paddies, sparking friction between farmers and herders. The hardest-hit areas are around Mopti and Tenenkou, Yongari-Mangari to the south, and along an imaginary line from Mourra to Sorme through Togoro Kotia. At that point, in other words, a quarter of the bourgoutières had been cleared. This phenomenon has not hit every area homogeneously. Clearance rates, by the yardstick used in today’s rural communes, can swing from 0% in one area to 82% in another.

60The most worrying cases are in communes around Tenenkou, south of Kewa, and in communes around Mopti (Socoura and Sio mainly). This is worrying for two reasons: first, it means herders are losing substantial amounts of grassland; second, it generates conflicts.

61Herding “infrastructure” (trails and shelters) is also under serious threat. In 1989, herders could not access 242 of the delta’s 1,014 shelters because of plantations. In other words, about 24% of the delta’s shelters were planted over. That, incidentally, was about how much bourgoutière had been cleared.

62A little later in the season, when the low-water stage came around (in the hot season), other issues surfaced emerged. Small village irrigation areas were “on” or “near” the paths that herds followed to water (which, of course meant that cattle trod on intensive crops).

63Settling these problems is a constant worry for territorial authorities and, today, for communal councils.


Farming schemes could jeopardise fish production on the plains

64Controlled-submersion rice plantations (ORM-type) and, to a lesser extent, village irrigation areas, are at issue here.

65Fish need unobstructed access (or at least unobstructed paths) to different areas in the water environment they dwell in (at least during the swell) to breed (Benech, 2002). Almost all fish species need to access the plains to breed, feed and grow. The proliferation of facilities (small dams and gate structures) by the river (to postpone paddy and bed flooding) obstruct access and end up denaturing the flooding-plain environment. Many species are scared of venturing into the narrow whirling courses created by the valves and spillways, and end up deserting these areas. Only a few species such as the Tilapia zillii can thrive on those plains. However, as they can damage crops, rice farmers build barriers along the small internal canals to block them (understandably, after all).

66Unchecked, these efforts to turn plains into controlled-submersion rice paddies could well hamper fish production across the delta. But there are schemes that allow rice and fish production to coexist (and indeed create synergies). These schemes are used extensively in Asia, but have not yet attracted enough attention in Africa’s rice-growing areas (especially in the Sahel region) in spite of their self-evident potential (Cofad, 2002).

Threats to nation-wide food security

67The population increase will also spell considerable changes in population structure (cf. “Large hydraulic schemes” p. 184). In 30 years’ time, about 50% of the population will live in cities. This means that the 50% of the population living in rural areas will have to produce enough food for themselves and for the other half. As unpredictable weather patterns hamper rain-fed crop yields in general and cereal production (millet) in particular, this is an issue.

68That is why the government has focused on boosting rice production (which “took off” in the late 1980s). To do so, authorities have put water management at the core of its agricultural-development agenda (MDRE, 1999). This strategy, SNDI (for Stratégie nationale de développement de l’irrigation/national irrigation-development strategy) involves building controlled-irrigation schemes for an extra 50,000 ha by 2007, plus shallow-water and controlled-submersion schemes for another 14,000 ha. By 2025, the goal is to extend the Office du Niger’s irrigated land area by 120,000 ha, to a total 200,000 ha.

69Even if these gravity-governed schemes were radically more efficient, irrigating the new areas would still involve diverting more water from the river during high-water and low-water stages alike. This would take us back to the critical scenario outlined above, involving the need to build a new storage/discharge dam upstream to support low-water flow. This would take an extra toll on the swell (especially when it reaches the delta) and therefore shrink flooded areas further. The risk, in other words, is that boosting rice production could tax meat and fish production (and fish production only barely covers the population’s needs today).

70Lastly, developments in the world rice market are a factor that warrants consideration in farming-development policies, especially when arbitrating between the various options. As a reminder, one of this market’s features today is the slight surplus coming from Asia. But this could change in coming years: Asia’s population will continue to increase, available farmland continue to shrink (from 0.15 to 0.09 ha per person), and the green revolution should wane (Trebuil and Hossain, 2004). Paradoxically, this scenario would be good news for Mali, as it would help its rice trade develop.


NBA issues

71Using the Niger River’s water more efficiently involves arbitrating between its various users and, first of all, between the countries skirting it. These decisions can only be made through ad hoc institutions. The idea of establishing a public organisation to improve and develop the Niger River basin is not new: it dates back to the early 1950s (the end of the colonial period), when MEAN (Mission d’étude et d’aménagement du Niger/Niger River research and improvement agency) was founded. This organisation controlled the river from its sources in Guinea to the Nigerian border. The agency that replaced MEAN, the CFN (Commission du fleuve Niger/ Niger River commission) was founded by the 1964 Act of Niamey (an agreement on navigation and economic cooperation between countries on the basin). This agreement showed that the Niger Basin’s new countries ranked river infrastructure and water management high on their agendas. Results, however, did not live up to expectations. Besides a 1970 study on the river’s navigability, the CFN achieved little. Its human and material resources never matched the mandates and colossal hopes that had been invested in it. In 1980, 17 years after its foundation, the CFN was replaced by the agency that is regarded as the basin authority today.

72NBA has had a difficult time. High hopes surrounded its inception, but its financial situation had become nothing short of catastrophic by 1986 (most member states had stopped paying their dues). But financial problems are only part of the NBA’s long crisis. Institutional issues explain the rest. The ties that shaped cooperation between the member states and dictated this international organisation’s mandates have been through several changes. Recurring shifts in cooperation patterns or strategic visions, in other words, have deprived the NBA of clear guidelines. It is also conceivable that the very notion of an international organisation in charge of harmonising water rights across the basin clashed with reticence on the part of States concerned about relinquishing their sovereignty. Funding agencies lost interest when they realised that member-state support was wilting. Geographically speaking, this basin is marginal in certain countries, explaining why some countries feel more involved than others.

73The NBA’s lasting difficulties also explain some of the mushrooming organisations and bilateral projects that ensued. The Ghenis project (Mali and Guinea) or the mixed Mali-Niger commission (a forum to harmonise the interests of these States and, in particular, to iron out the issues associated with low-water stages) are two examples.

74When they met in Abuja in October 1998, heads of State and funding agencies agreed to give the NBA a clean slate and a fresh start. They agreed on a three-year plan (spanning 2000 to 2002) to cement its organisation and bolster its human potential, and on a five-year action plan (starting in 2003) involving a number of new projects. These new projects include reviving the documentation centre, updating the hydrological databases (Niger-Hycos), and upgrading or creating a new river-flow model spanning the entire basin and a model to simulate water requirements.

75It is true that efficient NBA operation is a sine qua non for solving the geopolitical problems associated with using the basin’s waters. The Niger River’s upper and middle reaches are still relatively undeveloped. So the issues that might spark conflict between countries alongside those stretches are not yet manifest. But plans to build dams in Fomi (Guinea), Taoussa (Mali) and Kandadji (Niger), to list only the largest, will change this soon. If these three dams are built, it will be impossible to operate them independently. Moreover, funding-agency decisions seem tied to preliminary surveys ascertaining the impact that each of these dams will have on the river’s flow and ecosystems, and on the complex interactions between them.

76Consultation, it follows, hinges on efficient NBA operation and on strong ties with relevant national organisations. It also involves clarifying prerogatives and national-organisation powers.

A complex national institutional framework

77The fact that the philosophy underlying Mali’s water-management policies recently evolved possibly explains why the institutional framework is complex. Mali indeed recently switched from sector-based and technocratic management to integrated tributary-basin management (through IWRM). So it is fair to say that it is in a transition phase at this point. In particular, this process involves setting up new institutions without replacing existing ones (which is not necessarily desirable in all cases).

78To mention only one example, the recently-established Agence du Bassin du fleuve Niger (Niger River Basin agency) is run by the Ministry of the Environment. Water management, however, is still in the hands of the Ministry of Mines and Water, via the DNH (which runs the basin committees). Meaning that different authorities manage the Niger River’s basin (and sub-basins) in Mali.

79IWRM implementation (with decentralisation in the background) has opened the door to involving society in water management. Committees have mushroomed as a result. These committees often only have scant resources and their views generally only carry advisory weight. It is also fair to wonder about the rapport between communities with little human and material resources and “heavyweight” State organisations with limited but nonetheless real wherewithal, power over resources and technical expertise.

The difficulties for some groups to adjust to the new decentralisation-related “rules of the game”

80Farmland, grassland and fishing-area management has been handed over to territorial communities by law. But traditional authorities still hold dominant de facto authority. This can cause conflict between municipal and village authorities. Resource-management matters are referred, as appropriate, to the Dugukolotigi (master of land), Jitigi (master of water), Kungotigi (master of the bush), Baba Aougal (master of fishing) or Jowro (master of pasture).

81This decentralisation process, it is also worth pointing out, benefits sedentary communities (primarily farmers, in other words). This is because of the following:

  • decentralisation grants more power to communes, therefore to mayors. Mayors are often members of a commune’s sedentary community and therefore, by tradition, often farmers or members of the commune’s farmer community. Herders, for example, are only marginally represented, either because they are itinerant or because old political and territorial structures (the inland delta’s leyde) have made them reluctant to take part in communal electoral processes;
  • farming only takes up limited amounts of space. Livestock farming and fishing, conversely, cover vast areas (because of the itinerant element). The way in which decentralisation has distributed authority benefits sedentary farming more than itinerant livestock farming and fishing;
  • it is also fair to raise the question of the decentralisation drive’s grass-land management in the Niger River’s inland delta. This policy involves charging a herders a levy for using grassland whereas herders themselves – or at least their representatives (the jowros) – charged (and may still charge) their foreign peers a similar tax. The fact that, in the inland delta, herders and, to a lesser extent, fishers (migrating fishers, in particular) are noticeably underrepresented in democratically-elected bodies, can cause problems in these communities. And force them to take an in-depth look at their social and economic attitudes.

82The case of migrating fishers (Kassibo, 2000) proves this point. Fishing, as we have said, necessarily involves a sizeable itinerant element (in the case of its more professional component). Fishers have to fish a lot and all the time to cover the costs of their equipment and to make a profit. That means that they have to be where the fish is. But, as opposed to herders, fishers do not follow ancestral migratory patterns: they move less methodically, if nothing else because certain water ecosystems dry up and other appear (as in the case of Selingue). That is why fisher migrations do not necessarily involve return trips: fishers often settle in the areas they migrate to. However, fishers retain their “migrant” status and live as foreigners for a long time, even though they spend ten months out of twelve in their new place of abode. From an administrative standpoint, they would rather keep their ties to the area they come from instead of registering to vote or paying taxes in their host commune. As the decentralisation drive is moving natural resources, tax collection and public service into the hands of the smallest subdivisions of territorial communities (circles and communes), migrant fishers will soon be up against a wall. They will have to choose between becoming relegated, “second-class” citizens where they live (i.e. pay taxes and live by rules set by the “autochthons” without being able to access public services) or register in their host communities to achieve the same status as autochthons (Morand and Breuil, 2002). If they choose the second option, they will be able to take part in democratically-elected bodies (fishing councils, especially) in their host circles and communes, but will have to see their business as a source of income for the community. In return, they will be more encouraged to use the community’s public schools and maternity hospitals. Their economic and social ties to the village they came from (which they cling to today) will obviously suffer. As will their chances of moving on later (once they have invested themselves in the host community’s social and economic life). In other words, migrant fisher communities will have to alter their lifestyle social culture considerably.

Table des illustrations

Légende Source: Marie, 2002Figure 11. Flooded areas in the delta (up to lake Debo) paired with swell levels in Mopti
Fichier image/jpeg, 66k

© IRD Éditions, 2007

Conditions d’utilisation :